[Bio] / FigKernelPackages / Tracer.pm Repository:
ViewVC logotype

Diff of /FigKernelPackages/Tracer.pm

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.52, Wed Jun 14 01:12:42 2006 UTC revision 1.104, Fri May 9 06:22:49 2008 UTC
# Line 1  Line 1 
1  #  # -*- perl -*-
2    ########################################################################
3  # Copyright (c) 2003-2006 University of Chicago and Fellowship  # Copyright (c) 2003-2006 University of Chicago and Fellowship
4  # for Interpretations of Genomes. All Rights Reserved.  # for Interpretations of Genomes. All Rights Reserved.
5  #  #
# Line 13  Line 14 
14  # at info@ci.uchicago.edu or the Fellowship for Interpretation of  # at info@ci.uchicago.edu or the Fellowship for Interpretation of
15  # Genomes at veronika@thefig.info or download a copy from  # Genomes at veronika@thefig.info or download a copy from
16  # http://www.theseed.org/LICENSE.TXT.  # http://www.theseed.org/LICENSE.TXT.
17  #  ########################################################################
18    
19  package Tracer;  package Tracer;
20    
21      require Exporter;      require Exporter;
22      @ISA = ('Exporter');      @ISA = ('Exporter');
23      @EXPORT = qw(Trace T TSetup QTrace Confess Cluck Min Max Assert Open OpenDir TICK StandardSetup ScriptSetup ScriptFinish Insure ChDir);      @EXPORT = qw(Trace T TSetup QTrace Confess Cluck Min Max Assert Open OpenDir TICK StandardSetup EmergencyKey ETracing Constrain Insure ChDir Emergency Warn);
24      @EXPORT_OK = qw(GetFile GetOptions Merge MergeOptions ParseCommand ParseRecord UnEscape Escape);      @EXPORT_OK = qw(GetFile GetOptions Merge MergeOptions ParseCommand ParseRecord UnEscape Escape PrintLine PutLine);
25      use strict;      use strict;
26      use Carp qw(longmess croak);      use Carp qw(longmess croak carp);
27      use CGI;      use CGI;
28      use Cwd;      use Cwd;
29      use FIG_Config;      use FIG_Config;
# Line 31  Line 32 
32      use File::Basename;      use File::Basename;
33      use File::Path;      use File::Path;
34      use File::stat;      use File::stat;
35        use LWP::UserAgent;
36        use Time::HiRes 'gettimeofday';
37        use URI::Escape;
38        use Time::Local;
39        use POSIX qw(strftime);
40        use Time::Zone;
41    
42    
43  =head1 Tracing and Debugging Helpers  =head1 Tracing and Debugging Helpers
44    
45  =head2 Introduction  =head2 Tracing
46    
47  This package provides simple tracing for debugging and reporting purposes. To use it simply call the  This package provides simple tracing for debugging and reporting purposes. To use it simply call the
48  L</TSetup> method to set the options and call L</Trace> to write out trace messages. Each trace  L</TSetup> or L</ETracing> method to set the options and call L</Trace> to write out trace messages.
49  message has a I<trace level> and I<category> associated with it. In addition, the tracing package itself  L</TSetup> and L</ETracing> both establish a I<trace level> and a list of I<categories>. Similarly,
50  has a list of categories and a single trace level set by the B<TSetup> method. Only messages whose trace  each trace message has a I<trace level> and I<category> associated with it. Only messages whose trace
51  level is less than or equal to this package's trace level and whose category is activated will  level is less than or equal to the setup trace level and whose category is activated will
52  be written. Thus, a higher trace level on a message indicates that the message  be written. Thus, a higher trace level on a message indicates that the message
53  is less likely to be seen. A higher trace level passed to B<TSetup> means more trace messages will  is less likely to be seen, while a higher trace level passed to B<TSetup> means more trace messages will
54  appear. To generate a trace message, use the following syntax.  appear.
55    
56    =head3 Putting Trace Messages in Your Code
57    
58  C<< Trace($message) if T(errors => 4); >>  To generate a trace message, use the following syntax.
59    
60        Trace($message) if T(errors => 4);
61    
62  This statement will produce a trace message if the trace level is 4 or more and the C<errors>  This statement will produce a trace message if the trace level is 4 or more and the C<errors>
63  category is active. Note that the special category C<main> is always active, so  category is active. There is a special category C<main> that is always active, so
64    
65  C<< Trace($message) if T(main => 4); >>      Trace($message) if T(main => 4);
66    
67  will trace if the trace level is 4 or more.  will trace if the trace level is 4 or more.
68    
# Line 58  Line 70 
70  following call is made in the B<Sprout> package, it will appear if the C<Sprout> category is  following call is made in the B<Sprout> package, it will appear if the C<Sprout> category is
71  active and the trace level is 2 or more.  active and the trace level is 2 or more.
72    
73  C<< Trace($message) if T(2); >>      Trace($message) if T(2);
74    
75    In scripts, where no package name is available, the category defaults to C<main>.
76    
77    =head3 Custom Tracing
78    
79    Many programs have customized tracing configured using the L</TSetup> method. This is no longer
80    the preferred method, but a knowledge of how custom tracing works can make the more modern
81    L</Emergency Tracing> easier to understand.
82    
83  To set up tracing, you call the L</TSetup> method. The method takes as input a trace level, a list  To set up custom tracing, you call the L</TSetup> method. The method takes as input a trace level,
84  of category names, and a set of options. The trace level and list of category names are  a list of category names, and a destination. The trace level and list of category names are
85  specified as a space-delimited string. Thus  specified as a space-delimited string. Thus
86    
87  C<< TSetup('3 errors Sprout ERDB', 'HTML'); >>      TSetup('3 errors Sprout ERDB', 'TEXT');
88    
89  sets the trace level to 3, activates the C<errors>, C<Sprout>, and C<ERDB> categories, and  sets the trace level to 3, activates the C<errors>, C<Sprout>, and C<ERDB> categories, and
90  specifies that messages should be output as HTML paragraphs.  specifies that messages should be sent to the standard output.
91    
92  To turn on tracing for ALL categories, use an asterisk. The call below sets every category to  To turn on tracing for ALL categories, use an asterisk. The call below sets every category to
93  level 3 and writes the output to the standard error output. This sort of thing might be  level 3 and writes the output to the standard error output. This sort of thing might be
94  useful in a CGI environment.  useful in a CGI environment.
95    
96  C<< TSetup('3 *', 'WARN'); >>      TSetup('3 *', 'WARN');
97    
98  In addition to HTML and file output for trace messages, you can specify that the trace messages  In addition standard error and file output for trace messages, you can specify that the trace messages
99  be queued. The messages can then be retrieved by calling the L</QTrace> method. This approach  be queued. The messages can then be retrieved by calling the L</QTrace> method. This approach
100  is useful if you are building a web page. Instead of having the trace messages interspersed with  is useful if you are building a web page. Instead of having the trace messages interspersed with
101  the page output, they can be gathered together and displayed at the end of the page. This makes  the page output, they can be gathered together and displayed at the end of the page. This makes
102  it easier to debug page formatting problems.  it easier to debug page formatting problems.
103    
104  Finally, you can specify that all trace messages be emitted as warnings.  Finally, you can specify that all trace messages be emitted to a file, or the standard output and
105    a file at the same time. To trace to a file, specify the filename with an output character in front
106    of it.
107    
108        TSetup('4 SQL', ">$fileName");
109    
110    To trace to the standard output and a file at the same time, put a C<+> in front of the angle
111    bracket.
112    
113        TSetup('3 *', "+>$fileName");
114    
115  The flexibility of tracing makes it superior to simple use of directives like C<die> and C<warn>.  The flexibility of tracing makes it superior to simple use of directives like C<die> and C<warn>.
116  Tracer calls can be left in the code with minimal overhead and then turned on only when needed.  Tracer calls can be left in the code with minimal overhead and then turned on only when needed.
117  Thus, debugging information is available and easily retrieved even when the application is  Thus, debugging information is available and easily retrieved even when the application is
118  being used out in the field.  being used out in the field.
119    
120    =head3 Trace Levels
121    
122  There is no hard and fast rule on how to use trace levels. The following is therefore only  There is no hard and fast rule on how to use trace levels. The following is therefore only
123  a suggestion.  a suggestion.
124    
# Line 118  Line 149 
149    
150  =back  =back
151    
152    The format of trace messages is important because some utilities analyze trace files.
153    There are three fields-- the time stamp, the category name, and the text.
154    The time stamp is between square brackets and the category name between angle brackets.
155    After the category name there is a colon (C<:>) followed by the message text.
156    If the square brackets or angle brackets are missing, then the trace management
157    utilities assume that they are encountering a set of pre-formatted lines.
158    
159    Note, however, that this formatting is done automatically by the tracing functions. You
160    only need to know about it if you want to parse a trace file.
161    
162    =head3 Emergency Tracing
163    
164    Sometimes, you need a way for tracing to happen automatically without putting parameters
165    in a form or on the command line. Emergency tracing does this. You invoke emergency tracing
166    from the debug form, which is accessed from the [[DebugConsole]]. Emergency tracing requires
167    that you specify a tracing key. For command-line tools, the key is
168    taken from the C<TRACING> environment variable. For web services, the key is taken from
169    a cookie. Either way, the key tells the tracing facility who you are, so that you control
170    the tracing in your environment without stepping on other users.
171    
172    The key can be anything you want. If you don't have a key, the C<SetPassword> page will
173    generate one for you.
174    
175    You can activate and de-activate emergency tracing from the debugging control panel, as
176    well as display the trace file itself.
177    
178    To enable emergency tracing in your code, call
179    
180        ETracing($cgi)
181    
182    from a web script and
183    
184        ETracing()
185    
186    from a command-line script.
187    
188    The web script will look for the tracing key in the cookies, and the command-line
189    script will look for it in the C<TRACING> environment variable. If you are
190    using the L</StandardSetup> method or a [[WebApplication]], emergency tracing
191    will be configured automatically.
192    
193  =cut  =cut
194    
195  # Declare the configuration variables.  # Declare the configuration variables.
196    
197  my $Destination = "NONE";   # Description of where to send the trace output.  my $Destination = "WARN";   # Description of where to send the trace output.
198  my $TeeFlag = 0;            # TRUE if output is going to a file and to the  my $TeeFlag = 0;            # TRUE if output is going to a file and to the
199                              # standard output                              # standard output
200  my %Categories = ( main => 1 );  my %Categories = ( main => 1 );
201                              # hash of active category names                              # hash of active category names
202    my @LevelNames = qw(error warn notice info detail);
203  my $TraceLevel = 0;         # trace level; a higher trace level produces more  my $TraceLevel = 0;         # trace level; a higher trace level produces more
204                              # messages                              # messages
205  my @Queue = ();             # queued list of trace messages.  my @Queue = ();             # queued list of trace messages.
206  my $LastCategory = "main";  # name of the last category interrogated  my $LastCategory = "main";  # name of the last category interrogated
207    my $LastLevel = 0;          # level of the last test call
208  my $SetupCount = 0;         # number of times TSetup called  my $SetupCount = 0;         # number of times TSetup called
209  my $AllTrace = 0;           # TRUE if we are tracing all categories.  my $AllTrace = 0;           # TRUE if we are tracing all categories.
210    my $SavedCGI;               # CGI object passed to ETracing
211    my $CommandLine;            # Command line passed to StandardSetup
212    umask 2;                    # Fix the damn umask so everything is group-writable.
213    
214    =head2 Tracing Methods
215    
216    =head3 Setups
217    
218        my $count = Tracer::Setups();
219    
220    Return the number of times L</TSetup> has been called.
221    
222    This method allows for the creation of conditional tracing setups where, for example, we
223    may want to set up tracing if nobody else has done it before us.
224    
225    =cut
226    
227  =head2 Public Methods  sub Setups {
228        return $SetupCount;
229    }
230    
231  =head3 TSetup  =head3 TSetup
232    
233  C<< TSetup($categoryList, $target); >>      TSetup($categoryList, $target);
234    
235  This method is used to specify the trace options. The options are stored as package data  This method is used to specify the trace options. The options are stored as package data
236  and interrogated by the L</Trace> and L</T> methods.  and interrogated by the L</Trace> and L</T> methods.
# Line 188  Line 280 
280          }          }
281      }      }
282      # Now we need to process the destination information. The most important special      # Now we need to process the destination information. The most important special
283      # cases are the single ">", which requires we clear the file first, and the      # case is when we're writing to a file. This is indicated by ">" (overwrite) and
284      # "+" prefix which indicates a double echo.      # ">>" (append). A leading "+" for either indicates that we are also writing to
285        # the standard output (tee mode).
286      if ($target =~ m/^\+?>>?/) {      if ($target =~ m/^\+?>>?/) {
287          if ($target =~ m/^\+/) {          if ($target =~ m/^\+/) {
288              $TeeFlag = 1;              $TeeFlag = 1;
289              $target = substr($target, 1);              $target = substr($target, 1);
290          }          }
291          if ($target =~ m/^>[^>]/) {          if ($target =~ m/^>[^>]/) {
292                # We need to initialize the file (which clears it).
293              open TRACEFILE, $target;              open TRACEFILE, $target;
294              print TRACEFILE Now() . " Tracing initialized.\n";              print TRACEFILE "[" . Now() . "] [notice] [Tracer] Tracing initialized.\n";
295              close TRACEFILE;              close TRACEFILE;
296                # Set to append mode now that the file has been cleared.
297              $Destination = ">$target";              $Destination = ">$target";
298          } else {          } else {
299              $Destination = $target;              $Destination = $target;
# Line 210  Line 305 
305      $SetupCount++;      $SetupCount++;
306  }  }
307    
308  =head3 StandardSetup  =head3 SetLevel
309    
310  C<< my ($options, @parameters) = StandardSetup(\@categories, \%options, $parmHelp, @ARGV); >>      Tracer::SetLevel($newLevel);
311    
312  This method performs standard command-line parsing and tracing setup. The return  Modify the trace level. A higher trace level will cause more messages to appear.
 values are a hash of the command-line options and a list of the positional  
 parameters. Tracing is automatically set up and the command-line options are  
 validated.  
313    
314  This is a complex method that does a lot of grunt work. The parameters can  =over 4
 be more easily understood, however, once they are examined individually.  
315    
316  The I<categories> parameter is the most obtuse. It is a reference to a list of  =item newLevel
 special-purpose tracing categories. Most tracing categories are PERL package  
 names. So, for example, if you wanted to turn on tracing inside the B<Sprout>,  
 B<ERDB>, and B<SproutLoad> packages, you would specify the categories  
317    
318      ["Sprout", "SproutLoad", "ERDB"]  Proposed new trace level.
319    
320  This would cause trace messages in the specified three packages to appear in  =back
 the output. There are threer special tracing categories that are automatically  
 handled by this method. In other words, if you used L</TSetup> you would need  
 to include these categories manually, but if you use this method they are turned  
 on automatically.  
321    
322  =over 4  =cut
323    
324  =item FIG  sub SetLevel {
325        $TraceLevel = $_[0];
326    }
327    
328  Turns on trace messages inside the B<FIG> package.  =head3 ParseDate
329    
330  =item SQL      my $time = Tracer::ParseDate($dateString);
331    
332  Traces SQL commands and activity.  Convert a date into a PERL time number. This method expects a date-like string
333    and parses it into a number. The string must be vaguely date-like or it will
334    return an undefined value. Our requirement is that a month and day be
335    present and that three pieces of the date string (time of day, month and day,
336    year) be separated by likely delimiters, such as spaces, commas, and such-like.
337    
338  =item Tracer  If a time of day is present, it must be in military time with two digits for
339    everything but the hour.
340    
341  Traces error messages and call stacks.  The year must be exactly four digits.
342    
343  =back  Additional stuff can be in the string. We presume it's time zones or weekdays or something
344    equally innocuous. This means, however, that a sufficiently long sentence with date-like
345    parts in it may be interpreted as a date. Hopefully this will not be a problem.
346    
347  C<SQL> is only turned on if the C<-sql> option is specified in the command line.  It should be guaranteed that this method will parse the output of the L</Now> function.
 The trace level is specified using the C<-trace> command-line option. For example,  
 the following command line for C<TransactFeatures> turns on SQL tracing and runs  
 all tracing at level 3.  
348    
349      TransactFeatures -trace=3 -sql register ../xacts IDs.tbl  The parameters are as follows.
350    
351  Standard tracing is output to the standard output and echoed to the file  =over 4
 C<trace>I<$$>C<.log> in the FIG temporary directory, where I<$$> is the  
 process ID. You can also specify the C<user> parameter to put a user ID  
 instead of a process ID in the trace file name. So, for example  
352    
353  The default trace level is 2. To get all messages, specify a trace level of 4.  =item dateString
 For a genome-by-genome update, use 3.  
354    
355      TransactFeatures -trace=3 -sql -user=Bruce register ../xacts IDs.tbl  The date string to convert.
356    
357  would send the trace output to C<traceBruce.log> in the temporary directory.  =item RETURN
358    
359  The I<options> parameter is a reference to a hash containing the command-line  Returns a PERL time, that is, a number of seconds since the epoch, or C<undef> if
360  options, their default values, and an explanation of what they mean. Command-line  the date string is invalid. A valid date string must contain a month and day.
 options may be in the form of switches or keywords. In the case of a switch, the  
 option value is 1 if it is specified and 0 if it is not specified. In the case  
 of a keyword, the value is separated from the option name by an equal sign. You  
 can see this last in the command-line example above.  
361    
362  You can specify a different default trace level by setting C<$options->{trace}>  =back
 prior to calling this method.  
363    
364  An example at this point would help. Consider, for example, the command-line utility  =cut
 C<TransactFeatures>. It accepts a list of positional parameters plus the options  
 C<safe>, C<noAlias>, C<start>, and C<tblFiles>. To start up this command, we execute  
 the following code.  
365    
366      my ($options, @parameters) = Tracer::StandardSetup(["DocUtils"],  # Universal month conversion table.
367                          { safe => [0, "use database transactions"],  use constant MONTHS => {    Jan =>  0, January   =>  0, '01' =>  0,  '1' =>  0,
368                            noAlias => [0, "do not expect aliases in CHANGE transactions"],                              Feb =>  1, February  =>  1, '02' =>  1,  '2' =>  1,
369                            start => [' ', "start with this genome"],                              Mar =>  2, March     =>  2, '03' =>  2,  '3' =>  2,
370                            tblFiles => [0, "output TBL files containing the corrected IDs"] },                              Apr =>  3, April     =>  3, '04' =>  3,  '4' =>  3,
371                          "command transactionDirectory IDfile",                              May =>  4, May       =>  4, '05' =>  4,  '5' =>  4,
372                        @ARGV);                              Jun =>  5, June      =>  5, '06' =>  5,  '6' =>  5,
373                                Jul =>  6, July      =>  6, '07' =>  6,  '7' =>  6,
374                                Aug =>  7, August    =>  7, '08' =>  7,  '8' =>  7,
375                                Sep =>  8, September =>  8, '09' =>  8,  '9' =>  8,
376                                Oct =>  9, October  =>   9, '10' =>  9,
377                                Nov => 10, November =>  10, '11' => 10,
378                                Dec => 11, December =>  11, '12' => 11
379                            };
380    
381    sub ParseDate {
382        # Get the parameters.
383        my ($dateString) = @_;
384        # Declare the return variable.
385        my $retVal;
386        # Find the month and day of month. There are two ways that can happen. We check for the
387        # numeric style first. That way, if the user's done something like "Sun 12/22", then we
388        # won't be fooled into thinking the month is Sunday.
389        if ($dateString =~ m#\b(\d{1,2})/(\d{1,2})\b# || $dateString =~ m#\b(\w+)\s(\d{1,2})\b#) {
390            my ($mon, $mday) = (MONTHS->{$1}, $2);
391            # Insist that the month and day are valid.
392            if (defined($mon) && $2 >= 1 && $2 <= 31) {
393                # Find the time.
394                my ($hour, $min, $sec) = (0, 0, 0);
395                if ($dateString =~ /\b(\d{1,2}):(\d{2}):(\d{2})\b/) {
396                    ($hour, $min, $sec) = ($1, $2, $3);
397                }
398                # Find the year.
399                my $year;
400                if ($dateString =~ /\b(\d{4})\b/) {
401                    $year = $1;
402                } else {
403                    # Get the default year, which is this one. Note we must convert it to
404                    # the four-digit value expected by "timelocal".
405                    (undef, undef, undef, undef, undef, $year) = localtime();
406                    $year += 1900;
407                }
408                $retVal = timelocal($sec, $min, $hour, $mday, $mon, $year);
409            }
410        }
411        # Return the result.
412        return $retVal;
413    }
414    
415  The call to C<ParseCommand> specifies the default values for the options and  =head3 LogErrors
 stores the actual options in a hash that is returned as C<$options>. The  
 positional parameters are returned in C<@parameters>.  
416    
417  The following is a sample command line for C<TransactFeatures>.      Tracer::LogErrors($fileName);
418    
419      TransactFeatures -trace=2 -noAlias register ../xacts IDs.tbl  Route the standard error output to a log file.
420    
421  In this case, C<register>, C<../xacts>, and C<IDs.tbl> are the positional  =over 4
 parameters, and would find themselves in I<@parameters> after executing the  
 above code fragment. The tracing would be set to level 2, and the categories  
 would be C<FIG>, C<Tracer>, and <DocUtils>. C<FIG> and C<Tracer> are standard,  
 and C<DocUtils> was included because it came in within the first parameter  
 to this method. The I<$options> hash would be  
422    
423      { trace => 2, sql => 0, safe => 0,  =item fileName
       noAlias => 1, start => ' ', tblFiles => 0 }  
424    
425  Use of C<StandardSetup> in this way provides a simple way of performing  Name of the file to receive the error output.
 standard tracing setup and command-line parsing. Note that the caller is  
 not even aware of the command-line switches C<-trace> and C<-sql>, which  
 are used by this method to control the tracing. If additional tracing features  
 need to be added in the future, they can be processed by this method without  
 upsetting the command-line utilities.  
426    
427  If the C<background> option is specified on the command line, then the  =back
 standard and error outputs will be directed to files in the temporary  
 directory, using the same suffix as the trace file. So, if the command  
 line specified  
428    
429      -user=Bruce -background  =cut
430    
431  then the trace output would go to C<traceBruce.log>, the standard output to  sub LogErrors {
432  C<outBruce.log>, and the error output to C<errBruce.log>. This is designed to      # Get the file name.
433  simplify starting a command in the background.      my ($fileName) = @_;
434        # Open the file as the standard error output.
435        open STDERR, '>', $fileName;
436    }
437    
438  Finally, if the special option C<-h> is specified, the option names will  =head3 Trace
 be traced at level 0 and the program will exit without processing.  
 This provides a limited help capability. For example, if the user enters  
439    
440      TransactFeatures -h      Trace($message);
441    
442  he would see the following output.  Write a trace message to the target location specified in L</TSetup>. If there has not been
443    any prior call to B<TSetup>.
444    
445      TransactFeatures [options] command transactionDirectory IDfile  =over 4
         -trace    tracing level (default 2)  
         -sql      trace SQL commands  
         -safe     use database transactions  
         -noAlias  do not expect aliases in CHANGE transactions  
         -start    start with this genome  
         -tblFiles output TBL files containing the corrected IDs  
446    
447  The caller has the option of modifying the tracing scheme by placing a value  =item message
 for C<trace> in the incoming options hash. The default value can be overridden,  
 or the tracing to the standard output can be turned off by suffixing a minus  
 sign to the trace level. So, for example,  
448    
449      { trace => [0, "tracing level (default 0)"],  Message to write.
        ...  
450    
451  would set the default trace level to 0 instead of 2, while  =back
452    
453      { trace => ["2-", "tracing level (default 2)"],  =cut
        ...  
454    
455  would leave the default at 2, but trace only to the log file, not to the  sub Trace {
456  standard output.      # Get the parameters.
457        my ($message) = @_;
458        # Strip off any line terminators at the end of the message. We will add
459        # new-line stuff ourselves.
460        my $stripped = Strip($message);
461        # Compute the caller information.
462        my ($callPackage, $callFile, $callLine) = caller();
463        my $callFileTitle = basename($callFile);
464        # Check the caller.
465        my $callerInfo = ($callFileTitle ne "Tracer.pm" ? " [$callFileTitle $callLine]" : "");
466        # Get the timestamp.
467        my $timeStamp = Now();
468        # Build the prefix.
469        my $level = $LevelNames[$LastLevel] || "($LastLevel)";
470        my $prefix = "[$timeStamp] [$level] [$LastCategory]$callerInfo";
471        # Format the message.
472        my $formatted = "$prefix $stripped";
473        # Process according to the destination.
474        if ($Destination eq "TEXT") {
475            # Write the message to the standard output.
476            print "$formatted\n";
477        } elsif ($Destination eq "ERROR") {
478            # Write the message to the error output. Here, we want our prefix fields.
479            print STDERR "$formatted\n";
480        } elsif ($Destination eq "WARN") {
481            # Emit the message to the standard error output. It is presumed that the
482            # error logger will add its own prefix fields, the notable exception being
483            # the caller info.
484            print STDERR "$callerInfo$stripped\n";
485        } elsif ($Destination eq "QUEUE") {
486            # Push the message into the queue.
487            push @Queue, "$formatted";
488        } elsif ($Destination eq "HTML") {
489            # Convert the message to HTML and write it to the standard output.
490            my $escapedMessage = CGI::escapeHTML($stripped);
491            print "<p>$timeStamp $LastCategory $LastLevel: $escapedMessage</p>\n";
492        } elsif ($Destination =~ m/^>>/) {
493            # Write the trace message to an output file.
494            open(TRACING, $Destination) || die "Tracing open for \"$Destination\" failed: $!";
495            print TRACING "$formatted\n";
496            close TRACING;
497            # If the Tee flag is on, echo it to the standard output.
498            if ($TeeFlag) {
499                print "$formatted\n";
500            }
501        }
502    }
503    
504  The parameters to this method are as follows.  =head3 T
505    
506  =over 4      my $switch = T($category, $traceLevel);
507    
508  =item categories      or
509    
510  Reference to a list of tracing category names. These should be names of      my $switch = T($traceLevel);
 packages whose internal workings will need to be debugged to get the  
 command working.  
511    
512  =item options  Return TRUE if the trace level is at or above a specified value and the specified category
513    is active, else FALSE. If no category is specified, the caller's package name is used.
514    
515  Reference to a hash containing the legal options for the current command mapped  =over 4
 to their default values and descriptions. The user can override the defaults  
 by specifying the options as command-line switches prefixed by a hyphen.  
 Tracing-related options may be added to this hash. If the C<-h> option is  
 specified on the command line, the option descriptions will be used to  
 explain the options. To turn off tracing to the standard output, add a  
 minus sign to the value for C<trace> (see above).  
516    
517  =item parmHelp  =item category
518    
519  A string that vaguely describes the positional parameters. This is used  Category to which the message belongs. If not specified, the caller's package name is
520  if the user specifies the C<-h> option.  used.
521    
522  =item argv  =item traceLevel
523    
524  List of command line parameters, including the option switches, which must  Relevant tracing level.
 precede the positional parameters and be prefixed by a hyphen.  
525    
526  =item RETURN  =item RETURN
527    
528  Returns a list. The first element of the list is the reference to a hash that  TRUE if a message at the specified trace level would appear in the trace, else FALSE.
 maps the command-line option switches to their values. These will either be the  
 default values or overrides specified on the command line. The remaining  
 elements of the list are the position parameters, in order.  
529    
530  =back  =back
531    
532  =cut  =cut
533    
534  sub StandardSetup {  sub T {
535        # Declare the return variable.
536        my $retVal = 0;
537        # Only proceed if tracing is turned on.
538        if ($Destination ne "NONE") {
539      # Get the parameters.      # Get the parameters.
540      my ($categories, $options, $parmHelp, @argv) = @_;          my ($category, $traceLevel) = @_;
541      # Add the tracing options.          if (!defined $traceLevel) {
542      if (! exists $options->{trace}) {              # Here we have no category, so we need to get the calling package.
543          $options->{trace} = [2, "tracing level"];              # The calling package is normally the first parameter. If it is
544      }              # omitted, the first parameter will be the tracelevel. So, the
545      $options->{sql} = [0, "turn on SQL tracing"];              # first thing we do is shift the so-called category into the
546      $options->{h} = [0, "display command-line options"];              # $traceLevel variable where it belongs.
547      $options->{user} = [$$, "trace log file name suffix"];              $traceLevel = $category;
548      $options->{background} = [0, "spool standard and error output"];              my ($package, $fileName, $line) = caller;
549      # Create a parsing hash from the options hash. The parsing hash              # If there is no calling package, we default to "main".
550      # contains the default values rather than the default value              if (!$package) {
551      # and the description. While we're at it, we'll memorize the                  $category = "main";
552      # length of the longest option name.              } else {
553      my $longestName = 0;                  my @cats = split /::/, $package;
554                    $category = $cats[$#cats];
555                }
556            }
557            # Save the category name and level.
558            $LastCategory = $category;
559            $LastLevel = $traceLevel;
560            # Convert it to lower case before we hash it.
561            $category = lc $category;
562            # Validate the trace level.
563            if (ref $traceLevel) {
564                Confess("Bad trace level.");
565            } elsif (ref $TraceLevel) {
566                Confess("Bad trace config.");
567            }
568            # Make the check. Note that level 0 shows even if the category is turned off.
569            $retVal = ($traceLevel <= $TraceLevel && ($traceLevel == 0 || $AllTrace || exists $Categories{$category}));
570        }
571        # Return the computed result.
572        return $retVal;
573    }
574    
575    =head3 QTrace
576    
577        my $data = QTrace($format);
578    
579    Return the queued trace data in the specified format.
580    
581    =over 4
582    
583    =item format
584    
585    C<html> to format the data as an HTML list, C<text> to format it as straight text.
586    
587    =back
588    
589    =cut
590    
591    sub QTrace {
592        # Get the parameter.
593        my ($format) = @_;
594        # Create the return variable.
595        my $retVal = "";
596        # Only proceed if there is an actual queue.
597        if (@Queue) {
598            # Process according to the format.
599            if ($format =~ m/^HTML$/i) {
600                # Convert the queue into an HTML list.
601                $retVal = "<ul>\n";
602                for my $line (@Queue) {
603                    my $escapedLine = CGI::escapeHTML($line);
604                    $retVal .= "<li>$escapedLine</li>\n";
605                }
606                $retVal .= "</ul>\n";
607            } elsif ($format =~ m/^TEXT$/i) {
608                # Convert the queue into a list of text lines.
609                $retVal = join("\n", @Queue) . "\n";
610            }
611            # Clear the queue.
612            @Queue = ();
613        }
614        # Return the formatted list.
615        return $retVal;
616    }
617    
618    =head3 Confess
619    
620        Confess($message);
621    
622    Trace the call stack and abort the program with the specified message. When used with
623    the OR operator and the L</Assert> method, B<Confess> can function as a debugging assert.
624    So, for example
625    
626        Assert($recNum >= 0) || Confess("Invalid record number $recNum.");
627    
628    Will abort the program with a stack trace if the value of C<$recNum> is negative.
629    
630    =over 4
631    
632    =item message
633    
634    Message to include in the trace.
635    
636    =back
637    
638    =cut
639    
640    sub Confess {
641        # Get the parameters.
642        my ($message) = @_;
643        # Set up the category and level.
644        $LastCategory = "(confess)";
645        $LastLevel = 0;
646        if (! defined($FIG_Config::no_tool_hdr)) {
647            # Here we have a tool header. Display its length so that the user can adjust the line numbers.
648            my $toolHeaderFile = "$FIG_Config::fig_disk/dist/releases/current/$FIG_Config::arch/tool_hdr";
649            # Only proceed if the tool header file is actually present.
650            if (-f $toolHeaderFile) {
651                my $fh;
652                if (open $fh, "<$toolHeaderFile") {
653                    my @lines = <$fh>;
654                    Trace("Tool header has " . scalar(@lines) . " lines.");
655                }
656            }
657        }
658        # Trace the call stack.
659        Cluck($message);
660        # Abort the program.
661        croak(">>> $message");
662    }
663    
664    =head3 Warn
665    
666        Warn($message);
667    
668    This method traces an important message. If an RSS feed is configured
669    (via I<FIG_Config::error_feed>) and the tracing destination is C<WARN>,
670    then the message will be echoed to the feed. In general, a tracing
671    destination of C<WARN> indicates that the caller is running as a web
672    service in a production environment; however, this is not a requirement.
673    
674    To force warnings into the RSS feed even when the tracing destination
675    is not C<WARN>, simply specify the C<Feed> tracing module. This can be
676    configured automatically when L</StandardSetup> is used.
677    
678    The L</Cluck> method calls this one for its final message. Since
679    L</Confess> calls L</Cluck>, this means that any error which is caught
680    and confessed will put something in the feed. This insures that someone
681    will be alerted relatively quickly when a failure occurs.
682    
683    =over 4
684    
685    =item message
686    
687    Message to be traced.
688    
689    =back
690    
691    =cut
692    
693    sub Warn {
694        # Get the parameters.
695        my ($message) = @_;
696        # Trace the message.
697        Trace($message);
698        # Check for feed forcing.
699        my $forceFeed = exists $Categories{feed};
700        # An error here would be disastrous. Note, however, that we aren't too worried
701        # about losing events. The error log is always available for the occasions where
702        # we mess up. Note that if debug mode is specified, we do this stuff even in a
703        # test environment.
704        eval {
705            # Do we need to put this in the RSS feed?
706            if ($FIG_Config::error_feed && ($Destination eq 'WARN' || $forceFeed)) {
707                # Yes. We now need to compute the date, the link, and the title.
708                # First, the date, in a very specific format.
709                my $date = strftime("%a, %02e %b %H:%M:%S %Y ", localtime) .
710                    (tz_local_offset() / 30);
711                # Environment data goes in here. We start with the date.
712                my $environment = "$date.  ";
713                # If we need to recap the message (because it's too long to be a title), we'll
714                # put it in here.
715                my $recap;
716                # Copy the message and remove excess space.
717                my $title = $message;
718                $title =~ s/\s+/ /gs;
719                # If it's too long, we have to split it up.
720                if (length $title > 60) {
721                    # Put the full message in the environment string.
722                    $recap = $title;
723                    # Excerpt it as the title.
724                    $title = substr($title, 0, 50) . "...";
725                }
726                # If we have a CGI object, then this is a web error. Otherwise, it's
727                # command-line.
728                if (defined $SavedCGI) {
729                    # We're in a web service. The environment is the user's IP, and the link
730                    # is the URL that got us here.
731                    my $key = $ENV{HTTP_X_FORWARDED_FOR} || $ENV{REMOTE_ADDR};
732                    $environment .= "Event Reported at IP address $key.";
733                    my $url = $SavedCGI->url(-full => 1, -query => 1);
734                    # We need the user agent string and (if available) the referrer.
735                    # The referrer will be the link.
736                    $environment .= "User Agent $ENV{HTTP_USER_AGENT}";
737                    if ($ENV{HTTP_REFERER}) {
738                        my $link = $ENV{HTTP_REFERER};
739                        $environment .= " referred from <a href=\"$link\">$link</a>.";
740                    } else {
741                        $environment .= " referrer unknown.";
742                    }
743                    # Close off the sentence with the original link.
744                    $environment .= " URL of error is <a href=\"$url\">$url</a>.";
745                } else {
746                    # No CGI object, so we're a command-line tool. Use the tracing
747                    # key and the PID as the user identifier, and add the command.
748                    my $key = EmergencyKey();
749                    $environment .= "Event Reported by $key Process $$.";
750                    if ($CommandLine) {
751                        # We're in a StandardSetup script, so we have the real command line.
752                        $environment .= "\n<pre>" . CGI::escapeHTML($CommandLine) . "</pre>\n";
753                    } elsif ($ENV{_}) {
754                        # We're in a BASH script, so the command has been stored in the _ variable.
755                        $environment .= "  Command = " . CGI::escapeHTML($ENV{_}) . "\n";
756                    }
757                }
758                # Build a GUID. We use the current time, the title, and the process ID,
759                # then digest the result.
760                my $guid = Digest::MD5::md5_base64(gettimeofday(), $title, $$);
761                # Finally, the description. This is a stack trace plus various environmental stuff.
762                my $stackTrace = "";
763                my @trace = LongMess();
764                # Only proceed if we got something back.
765                if (scalar(@trace) > 0) {
766                    $trace[0] =~ s/Tracer::Warn.+?called/Event occurred/;
767                    $stackTrace = "Stack trace:<pre>" . join("\n", @trace, "</pre>");
768                }
769                # We got the stack trace. Now it's time to put it all together.
770                # We have a goofy thing here in that we need to HTML-escape some sections of the description
771                # twice. They will be escaped once here, and then once when written by XML::Simple. They are
772                # unescaped once when processed by the RSS reader, and stuff in the description is treated as
773                # HTML. So, anything escaped here is treated as a literal when viewed in the RSS reader, but
774                # our <br>s and <pre>s are used to format the description.
775                $recap = (defined $recap ? "<em>" . CGI::escapeHTML($recap) . "</em><br /><br />" : "");
776                my $description = "$recap$environment  $stackTrace";
777                # Okay, we have all the pieces. Create a hash of the new event.
778                my $newItem = { title => $title,
779                                description => $description,
780                                category => $LastCategory,
781                                pubDate => $date,
782                                guid => $guid,
783                               };
784                # We need XML capability for this.
785                require XML::Simple;
786                # The RSS document goes in here.
787                my $rss;
788                # Get the name of the RSS file. It's in the FIG temporary directory.
789                my $fileName = "$FIG_Config::temp/$FIG_Config::error_feed";
790                # Does it exist?
791                if (-s $fileName) {
792                    # Slurp it in.
793                    $rss = XML::Simple::XMLin($fileName, ForceArray => ['item']);
794                } else {
795                    my $size = -s $fileName;
796                    # Create an empty channel.
797                    $rss = {
798                        channel => {
799                            title => 'NMPDR Warning Feed',
800                            link => "$FIG_Config::temp_url/$FIG_Config::error_feed",
801                            description => "Important messages regarding the status of the NMPDR.",
802                            generator => "NMPDR Trace Facility",
803                            docs => "http://blogs.law.harvard.edu/tech/rss",
804                            item => []
805                        },
806                    };
807                }
808                # Get the channel object.
809                my $channel = $rss->{channel};
810                # Update the last-build date.
811                $channel->{lastBuildDate} = $date;
812                # Get the item array.
813                my $items = $channel->{item};
814                # Insure it has only 100 entries.
815                while (scalar @{$items} > 100) {
816                    pop @{$items};
817                }
818                # Add our new item at the front.
819                unshift @{$items}, $newItem;
820                # Create the XML. Note we do not include the root or the declaration. XML Simple can't handle
821                # the requirements for those.
822                my $xml = XML::Simple::XMLout($rss, NoAttr => 1, RootName => undef, XmlDecl => '');
823                # Here we put in the root and declaration. The problem is that the root has to have the version attribute
824                # in it. So, we suppress the root and do it by hand, and that requires suppressing the declaration, too.
825                $xml = "<?xml version=\"1.0\" encoding=\"utf-8\"?>\n<rss version=\"2.0\">$xml\n</rss>";
826                # We don't use Open here because we can't afford an error.
827                if (open XMLOUT, ">$fileName") {
828                    print XMLOUT $xml;
829                    close XMLOUT;
830                }
831            }
832        };
833        if ($@) {
834            # If the feed failed, we need to know why. The error will be traced, but this method will not be involved
835            # (which is a good thing).
836            my $error = $@;
837            Trace("Feed Error: $error") if T(Feed => 0);
838        }
839    }
840    
841    =head3 Assert
842    
843        Assert($condition1, $condition2, ... $conditionN);
844    
845    Return TRUE if all the conditions are true. This method can be used in conjunction with
846    the OR operator and the L</Confess> method as a debugging assert.
847    So, for example
848    
849        Assert($recNum >= 0) || Confess("Invalid record number $recNum.");
850    
851    Will abort the program with a stack trace if the value of C<$recNum> is negative.
852    
853    =cut
854    sub Assert {
855        my $retVal = 1;
856        LOOP: for my $condition (@_) {
857            if (! $condition) {
858                $retVal = 0;
859                last LOOP;
860            }
861        }
862        return $retVal;
863    }
864    
865    =head3 Cluck
866    
867        Cluck($message);
868    
869    Trace the call stack. Note that for best results, you should qualify the call with a
870    trace condition. For example,
871    
872        Cluck("Starting record parse.") if T(3);
873    
874    will only trace the stack if the trace level for the package is 3 or more.
875    
876    =over 4
877    
878    =item message
879    
880    Message to include in the trace.
881    
882    =back
883    
884    =cut
885    
886    sub Cluck {
887        # Get the parameters.
888        my ($message) = @_;
889        # Trace what's happening.
890        Trace("Stack trace for event: $message");
891        # Get the stack trace.
892        my @trace = LongMess();
893        # Convert the trace to a series of messages.
894        for my $line (@trace) {
895            # Replace the tab at the beginning with spaces.
896            $line =~ s/^\t/    /;
897            # Trace the line.
898            Trace($line);
899        }
900        # Issue a warning. This displays the event message and inserts it into the RSS error feed.
901        Warn($message);
902    }
903    
904    =head3 LongMess
905    
906        my @lines = Tracer::LongMess();
907    
908    Return a stack trace with all tracing methods removed. The return will be in the form of a list
909    of message strings.
910    
911    =cut
912    
913    sub LongMess {
914        # Declare the return variable.
915        my @retVal = ();
916        my $confession = longmess("");
917        for my $line (split /\s*\n/, $confession) {
918            unless ($line =~ /Tracer\.pm/) {
919                # Here we have a line worth keeping. Push it onto the result list.
920                push @retVal, $line;
921            }
922        }
923        # Return the result.
924        return @retVal;
925    }
926    
927    =head3 ScriptSetup (deprecated)
928    
929        my ($cgi, $varHash) = ScriptSetup($noTrace);
930    
931    Perform standard tracing and debugging setup for scripts. The value returned is
932    the CGI object followed by a pre-built variable hash. At the end of the script,
933    the client should call L</ScriptFinish> to output the web page.
934    
935    This method calls L</ETracing> to configure tracing, which allows the tracing
936    to be configured via the emergency tracing form on the debugging control panel.
937    Tracing will then be turned on automatically for all programs that use the L</ETracing>
938    method, which includes every program that uses this method or L</StandardSetup>.
939    
940    =over 4
941    
942    =item noTrace (optional)
943    
944    If specified, tracing will be suppressed. This is useful if the script wants to set up
945    tracing manually.
946    
947    =item RETURN
948    
949    Returns a two-element list consisting of a CGI query object and a variable hash for
950    the output page.
951    
952    =back
953    
954    =cut
955    
956    sub ScriptSetup {
957        # Get the parameters.
958        my ($noTrace) = @_;
959        # Get the CGI query object.
960        my $cgi = CGI->new();
961        # Set up tracing if it's not suppressed.
962        ETracing($cgi) unless $noTrace;
963        # Create the variable hash.
964        my $varHash = { results => '' };
965        # Return the query object and variable hash.
966        return ($cgi, $varHash);
967    }
968    
969    =head3 ETracing
970    
971        ETracing($parameter);
972    
973    Set up emergency tracing. Emergency tracing is tracing that is turned
974    on automatically for any program that calls this method. The emergency
975    tracing parameters are stored in a a file identified by a tracing key.
976    If this method is called with a CGI object, then the tracing key is
977    taken from a cookie. If it is called with no parameters, then the tracing
978    key is taken from an environment variable. If it is called with a string,
979    the tracing key is that string.
980    
981    =over 4
982    
983    =item parameter
984    
985    A parameter from which the tracing key is computed. If it is a scalar,
986    that scalar is used as the tracing key. If it is a CGI object, the
987    tracing key is taken from the C<IP> cookie. If it is omitted, the
988    tracing key is taken from the C<TRACING> environment variable. If it
989    is a CGI object and emergency tracing is not on, the C<Trace> and
990    C<TF> parameters will be used to determine the type of tracing.
991    
992    =back
993    
994    =cut
995    
996    sub ETracing {
997        # Get the parameter.
998        my ($parameter) = @_;
999        # Check for CGI mode.
1000        if (defined $parameter && ref $parameter eq 'CGI') {
1001            $SavedCGI = $parameter;
1002        } else {
1003            $SavedCGI = undef;
1004        }
1005        # Default to no tracing except errors.
1006        my ($tracing, $dest) = ("0", "WARN");
1007        # Check for emergency tracing.
1008        my $tkey = EmergencyKey($parameter);
1009        my $emergencyFile = EmergencyFileName($tkey);
1010        if (-e $emergencyFile) {
1011            # We have the file. Read in the data.
1012            my @tracing = GetFile($emergencyFile);
1013            # Pull off the time limit.
1014            my $expire = shift @tracing;
1015            # Convert it to seconds.
1016            $expire *= 3600;
1017            # Check the file data.
1018            my $stat = stat($emergencyFile);
1019            my ($now) = gettimeofday;
1020            if ($now - $stat->mtime > $expire) {
1021                # Delete the expired file.
1022                unlink $emergencyFile;
1023            } else {
1024                # Emergency tracing is on. Pull off the destination and
1025                # the trace level;
1026                $dest = shift @tracing;
1027                my $level = shift @tracing;
1028                # Convert the destination to a real tracing destination.
1029                # temp directory.
1030                $dest = EmergencyTracingDest($tkey, $dest);
1031                # Insure Tracer is specified.
1032                my %moduleHash = map { $_ => 1 } @tracing;
1033                $moduleHash{Tracer} = 1;
1034                # Set the trace parameter.
1035                $tracing = join(" ", $level, sort keys %moduleHash);
1036            }
1037        } elsif (defined $SavedCGI) {
1038            # There's no emergency tracing, but we have a CGI object, so check
1039            # for tracing from the form parameters.
1040            if ($SavedCGI->param('Trace')) {
1041                # Here the user has requested tracing via a form.
1042                $dest = ($SavedCGI->param('TF') ? ">$FIG_Config::temp/Trace$$.log" : "QUEUE");
1043                $tracing = $SavedCGI->param('Trace') . " Tracer";
1044            }
1045        }
1046        # Setup the tracing we've determined from all the stuff above.
1047        TSetup($tracing, $dest);
1048        # Check to see if we're a web script.
1049        if (defined $SavedCGI) {
1050            # Yes we are. Trace the form and environment data.
1051            TraceParms($SavedCGI);
1052            # Check for RAW mode. In raw mode, we print a fake header so that we see everything
1053            # emitted by the script in its raw form.
1054            if (T(Raw => 3)) {
1055                print CGI::header(-type => 'text/plain', -tracing => 'Raw');
1056            }
1057        }
1058    }
1059    
1060    =head3 EmergencyFileName
1061    
1062        my $fileName = Tracer::EmergencyFileName($tkey);
1063    
1064    Return the emergency tracing file name. This is the file that specifies
1065    the tracing information.
1066    
1067    =over 4
1068    
1069    =item tkey
1070    
1071    Tracing key for the current program.
1072    
1073    =item RETURN
1074    
1075    Returns the name of the file to contain the emergency tracing information.
1076    
1077    =back
1078    
1079    =cut
1080    
1081    sub EmergencyFileName {
1082        # Get the parameters.
1083        my ($tkey) = @_;
1084        # Compute the emergency tracing file name.
1085        return "$FIG_Config::temp/Emergency$tkey.txt";
1086    }
1087    
1088    =head3 EmergencyFileTarget
1089    
1090        my $fileName = Tracer::EmergencyFileTarget($tkey);
1091    
1092    Return the emergency tracing target file name. This is the file that receives
1093    the tracing output for file-based tracing.
1094    
1095    =over 4
1096    
1097    =item tkey
1098    
1099    Tracing key for the current program.
1100    
1101    =item RETURN
1102    
1103    Returns the name of the file to contain the trace output.
1104    
1105    =back
1106    
1107    =cut
1108    
1109    sub EmergencyFileTarget {
1110        # Get the parameters.
1111        my ($tkey) = @_;
1112        # Compute the emergency tracing file name.
1113        return "$FIG_Config::temp/trace$tkey.log";
1114    }
1115    
1116    =head3 EmergencyTracingDest
1117    
1118        my $dest = Tracer::EmergencyTracingDest($tkey, $myDest);
1119    
1120    This method converts an emergency tracing destination to a real
1121    tracing destination. The main difference is that if the
1122    destination is C<FILE> or C<APPEND>, we convert it to file
1123    output. If the destination is C<DUAL>, we convert it to file
1124    and standard output.
1125    
1126    =over 4
1127    
1128    =item tkey
1129    
1130    Tracing key for this environment.
1131    
1132    =item myDest
1133    
1134    Destination from the emergency tracing file.
1135    
1136    =item RETURN
1137    
1138    Returns a destination that can be passed into L</TSetup>.
1139    
1140    =back
1141    
1142    =cut
1143    
1144    sub EmergencyTracingDest {
1145        # Get the parameters.
1146        my ($tkey, $myDest) = @_;
1147        # Declare the return variable.
1148        my $retVal = $myDest;
1149        # Process according to the destination value.
1150        if ($myDest eq 'FILE') {
1151            $retVal = ">" . EmergencyFileTarget($tkey);
1152        } elsif ($myDest eq 'APPEND') {
1153            $retVal = ">>" . EmergencyFileTarget($tkey);
1154        } elsif ($myDest eq 'DUAL') {
1155            $retVal = "+>" . EmergencyFileTarget($tkey);
1156        } elsif ($myDest eq 'WARN') {
1157            $retVal = "WARN";
1158        }
1159        # Return the result.
1160        return $retVal;
1161    }
1162    
1163    =head3 Emergency
1164    
1165        Emergency($key, $hours, $dest, $level, @modules);
1166    
1167    Turn on emergency tracing. This method is normally invoked over the web from
1168    a debugging console, but it can also be called by the C<trace.pl> script.
1169    The caller specifies the duration of the emergency in hours, the desired tracing
1170    destination, the trace level, and a list of the trace modules to activate.
1171    For the length of the duration, when a program in an environment with the
1172    specified tracing key active invokes a Sprout CGI script, tracing will be
1173    turned on automatically. See L</TSetup> for more about tracing setup and
1174    L</ETracing> for more about emergency tracing.
1175    
1176    =over 4
1177    
1178    =item tkey
1179    
1180    The tracing key. This is used to identify the control file and the trace file.
1181    
1182    =item hours
1183    
1184    Number of hours to keep emergency tracing alive.
1185    
1186    =item dest
1187    
1188    Tracing destination. If no path information is specified for a file
1189    destination, it is put in the FIG temporary directory.
1190    
1191    =item level
1192    
1193    Tracing level. A higher level means more trace messages.
1194    
1195    =item modules
1196    
1197    A list of the tracing modules to activate.
1198    
1199    =back
1200    
1201    =cut
1202    
1203    sub Emergency {
1204        # Get the parameters.
1205        my ($tkey, $hours, $dest, $level, @modules) = @_;
1206        # Create the emergency file.
1207        my $specFile = EmergencyFileName($tkey);
1208        my $outHandle = Open(undef, ">$specFile");
1209        print $outHandle join("\n", $hours, $dest, $level, @modules, "");
1210    }
1211    
1212    =head3 EmergencyKey
1213    
1214        my $tkey = EmergencyKey($parameter);
1215    
1216    Return the Key to be used for emergency tracing. This could be an IP address,
1217     a session ID, or a user name, depending on the environment.
1218    
1219    =over 4
1220    
1221    =item parameter
1222    
1223    Parameter defining the method for finding the tracing key. If it is a scalar,
1224    then it is presumed to be the tracing key itself. If it is a CGI object, then
1225    the tracing key is taken from the C<IP> cookie. Otherwise, the tracing key is
1226    taken from the C<TRACING> environment variable.
1227    
1228    =item RETURN
1229    
1230    Returns the key to be used for labels in emergency tracing.
1231    
1232    =back
1233    
1234    =cut
1235    
1236    sub EmergencyKey {
1237        # Get the parameters.
1238        my ($parameter) = @_;
1239        # Declare the return variable.
1240        my $retVal;
1241        # Determine the parameter type.
1242        if (! defined $parameter) {
1243            # Here we're supposed to check the environment. If that fails, we
1244            # get the effective login ID.
1245            $retVal = $ENV{TRACING} || scalar getpwuid($<);
1246        } else {
1247            my $ptype = ref $parameter;
1248            if ($ptype eq 'CGI') {
1249                # Here we were invoked from a web page. Look for a cookie.
1250                $retVal = $parameter->cookie('IP');
1251            } elsif (! $ptype) {
1252                # Here the key was passed in.
1253                $retVal = $parameter;
1254            }
1255        }
1256        # If no luck finding a key, use the PID.
1257        if (! defined $retVal) {
1258            $retVal = $$;
1259        }
1260        # Return the result.
1261        return $retVal;
1262    }
1263    
1264    
1265    =head3 TraceParms
1266    
1267        Tracer::TraceParms($cgi);
1268    
1269    Trace the CGI parameters at trace level CGI => 3 and the environment variables
1270    at level CGI => 4. A self-referencing URL is traced at level CGI => 2.
1271    
1272    =over 4
1273    
1274    =item cgi
1275    
1276    CGI query object containing the parameters to trace.
1277    
1278    =back
1279    
1280    =cut
1281    
1282    sub TraceParms {
1283        # Get the parameters.
1284        my ($cgi) = @_;
1285        if (T(CGI => 2)) {
1286            # Here we trace the GET-style URL for the script.
1287            Trace("[URL] " . $cgi->url(-relative => 1, -query => 1));
1288        }
1289        if (T(CGI => 3)) {
1290            # Here we want to trace the parameter data.
1291            my @names = $cgi->param;
1292            for my $parmName (sort @names) {
1293                # Note we skip the Trace parameters, which are for our use only.
1294                if ($parmName ne 'Trace' && $parmName ne 'TF') {
1295                    my @values = $cgi->param($parmName);
1296                    Trace("[CGI] $parmName = " . join(", ", @values));
1297                }
1298            }
1299            # Display the request method.
1300            my $method = $cgi->request_method();
1301            Trace("Method: $method");
1302        }
1303        if (T(CGI => 4)) {
1304            # Here we want the environment data too.
1305            for my $envName (sort keys %ENV) {
1306                Trace("[ENV] $envName = $ENV{$envName}");
1307            }
1308        }
1309    }
1310    
1311    =head3 TraceImages
1312    
1313        Tracer::TraceImages($htmlString);
1314    
1315    Trace information about all of an html document's images. The tracing
1316    will be for type "IMG" at level 3. The image's source string
1317    will be displayed. This is generally either the URL of the image or
1318    raw data for the image itself. If the source is too long, only the first 300
1319    characters will be shown at trace level 3. The entire source will be shown,
1320    however, at trace level 4. This method is not very smart, and might catch
1321    Javascript code, but it is still useful when debugging the arcane
1322    behavior of images in multiple browser environments.
1323    
1324    =over 4
1325    
1326    =item htmlString
1327    
1328    HTML text for an outgoing web page.
1329    
1330    =back
1331    
1332    =cut
1333    
1334    sub TraceImages {
1335        # Only proceed if we're at the proper trace level.
1336        if (T(IMG => 3)) {
1337            # For performance reasons we're manipulating $_[0] instead of retrieving the string
1338            # into a variable called "$htmlString". This is because we expect html strings to be
1339            # long, and don't want to copy them any more than we have to.
1340            Trace(length($_[0]) . " characters in web page.");
1341            # Loop through the HTML, culling image tags.
1342            while ($_[0] =~ /<img\s+[^>]+?src="([^"]+)"/sgi) {
1343                # Extract the source string and determine whether or not it's too long.
1344                my $srcString = $1;
1345                my $pos = pos($_[0]) - length($srcString);
1346                my $excess = length($srcString) - 300;
1347                # We'll put the display string in here.
1348                my $srcDisplay = $srcString;
1349                # If it's a data string, split it at the comma.
1350                $srcDisplay =~ s/^(data[^,]+,)/$1\n/;
1351                # If there's no excess or we're at trace level 4, we're done. At level 3 with
1352                # a long string, however, we only show the first 300 characters.
1353                if ($excess > 0 && ! T(IMG => 4)) {
1354                    $srcDisplay = substr($srcDisplay,0,300) . "\nplus $excess characters.";
1355                }
1356                # Output the trace message.
1357                Trace("Image tag at position $pos:\n$srcDisplay");
1358            }
1359        }
1360    }
1361    
1362    
1363    =head3 ScriptFinish (deprecated)
1364    
1365        ScriptFinish($webData, $varHash);
1366    
1367    Output a web page at the end of a script. Either the string to be output or the
1368    name of a template file can be specified. If the second parameter is omitted,
1369    it is assumed we have a string to be output; otherwise, it is assumed we have the
1370    name of a template file. The template should have the variable C<DebugData>
1371    specified in any form that invokes a standard script. If debugging mode is turned
1372    on, a form field will be put in that allows the user to enter tracing data.
1373    Trace messages will be placed immediately before the terminal C<BODY> tag in
1374    the output, formatted as a list.
1375    
1376    A typical standard script would loook like the following.
1377    
1378        BEGIN {
1379            # Print the HTML header.
1380            print "CONTENT-TYPE: text/html\n\n";
1381        }
1382        use Tracer;
1383        use CGI;
1384        use FIG;
1385        # ... more uses ...
1386    
1387        my ($cgi, $varHash) = ScriptSetup();
1388        eval {
1389            # ... get data from $cgi, put it in $varHash ...
1390        };
1391        if ($@) {
1392            Trace("Script Error: $@") if T(0);
1393        }
1394        ScriptFinish("Html/MyTemplate.html", $varHash);
1395    
1396    The idea here is that even if the script fails, you'll see trace messages and
1397    useful output.
1398    
1399    =over 4
1400    
1401    =item webData
1402    
1403    A string containing either the full web page to be written to the output or the
1404    name of a template file from which the page is to be constructed. If the name
1405    of a template file is specified, then the second parameter must be present;
1406    otherwise, it must be absent.
1407    
1408    =item varHash (optional)
1409    
1410    If specified, then a reference to a hash mapping variable names for a template
1411    to their values. The template file will be read into memory, and variable markers
1412    will be replaced by data in this hash reference.
1413    
1414    =back
1415    
1416    =cut
1417    
1418    sub ScriptFinish {
1419        # Get the parameters.
1420        my ($webData, $varHash) = @_;
1421        # Check for a template file situation.
1422        my $outputString;
1423        if (defined $varHash) {
1424            # Here we have a template file. We need to determine the template type.
1425            my $template;
1426            if ($FIG_Config::template_url && $webData =~ /\.php$/) {
1427                $template = "$FIG_Config::template_url/$webData";
1428            } else {
1429                $template = "<<$webData";
1430            }
1431            $outputString = PageBuilder::Build($template, $varHash, "Html");
1432        } else {
1433            # Here the user gave us a raw string.
1434            $outputString = $webData;
1435        }
1436        # Check for trace messages.
1437        if ($Destination ne "NONE" && $TraceLevel > 0) {
1438            # We have trace messages, so we want to put them at the end of the body. This
1439            # is either at the end of the whole string or at the beginning of the BODY
1440            # end-tag.
1441            my $pos = length $outputString;
1442            if ($outputString =~ m#</body>#gi) {
1443                $pos = (pos $outputString) - 7;
1444            }
1445            # If the trace messages were queued, we unroll them. Otherwise, we display the
1446            # destination.
1447            my $traceHtml;
1448            if ($Destination eq "QUEUE") {
1449                $traceHtml = QTrace('Html');
1450            } elsif ($Destination =~ /^>>(.+)$/) {
1451                # Here the tracing output it to a file. We code it as a hyperlink so the user
1452                # can copy the file name into the clipboard easily.
1453                my $actualDest = $1;
1454                $traceHtml = "<p>Tracing output to $actualDest.</p>\n";
1455            } else {
1456                # Here we have one of the special destinations.
1457                $traceHtml = "<P>Tracing output type is $Destination.</p>\n";
1458            }
1459            substr $outputString, $pos, 0, $traceHtml;
1460        }
1461        # Write the output string.
1462        print $outputString;
1463    }
1464    
1465    =head2 Command-Line Utility Methods
1466    
1467    =head3 SendSMS
1468    
1469        my $msgID = Tracer::SendSMS($phoneNumber, $msg);
1470    
1471    Send a text message to a phone number using Clickatell. The FIG_Config file must contain the
1472    user name, password, and API ID for the relevant account in the hash reference variable
1473    I<$FIG_Config::phone>, using the keys C<user>, C<password>, and C<api_id>. For
1474    example, if the user name is C<BruceTheHumanPet>, the password is C<silly>, and the API ID
1475    is C<2561022>, then the FIG_Config file must contain
1476    
1477        $phone =  { user => 'BruceTheHumanPet',
1478                    password => 'silly',
1479                    api_id => '2561022' };
1480    
1481    The original purpose of this method was to insure Bruce would be notified immediately when the
1482    Sprout Load terminates. Care should be taken if you do not wish Bruce to be notified immediately
1483    when you call this method.
1484    
1485    The message ID will be returned if successful, and C<undef> if an error occurs.
1486    
1487    =over 4
1488    
1489    =item phoneNumber
1490    
1491    Phone number to receive the message, in international format. A United States phone number
1492    would be prefixed by "1". A British phone number would be prefixed by "44".
1493    
1494    =item msg
1495    
1496    Message to send to the specified phone.
1497    
1498    =item RETURN
1499    
1500    Returns the message ID if successful, and C<undef> if the message could not be sent.
1501    
1502    =back
1503    
1504    =cut
1505    
1506    sub SendSMS {
1507        # Get the parameters.
1508        my ($phoneNumber, $msg) = @_;
1509        # Declare the return variable. If we do not change it, C<undef> will be returned.
1510        my $retVal;
1511        # Only proceed if we have phone support.
1512        if (! defined $FIG_Config::phone) {
1513            Trace("Phone support not present in FIG_Config.") if T(1);
1514        } else {
1515            # Get the phone data.
1516            my $parms = $FIG_Config::phone;
1517            # Get the Clickatell URL.
1518            my $url = "http://api.clickatell.com/http/";
1519            # Create the user agent.
1520            my $ua = LWP::UserAgent->new;
1521            # Request a Clickatell session.
1522            my $resp = $ua->post("$url/sendmsg", { user => $parms->{user},
1523                                         password => $parms->{password},
1524                                         api_id => $parms->{api_id},
1525                                         to => $phoneNumber,
1526                                         text => $msg});
1527            # Check for an error.
1528            if (! $resp->is_success) {
1529                Trace("Alert failed.") if T(1);
1530            } else {
1531                # Get the message ID.
1532                my $rstring = $resp->content;
1533                if ($rstring =~ /^ID:\s+(.*)$/) {
1534                    $retVal = $1;
1535                } else {
1536                    Trace("Phone attempt failed with $rstring") if T(1);
1537                }
1538            }
1539        }
1540        # Return the result.
1541        return $retVal;
1542    }
1543    
1544    =head3 StandardSetup
1545    
1546        my ($options, @parameters) = StandardSetup(\@categories, \%options, $parmHelp, @ARGV);
1547    
1548    This method performs standard command-line parsing and tracing setup. The return
1549    values are a hash of the command-line options and a list of the positional
1550    parameters. Tracing is automatically set up and the command-line options are
1551    validated.
1552    
1553    This is a complex method that does a lot of grunt work. The parameters can
1554    be more easily understood, however, once they are examined individually.
1555    
1556    The I<categories> parameter is the most obtuse. It is a reference to a list of
1557    special-purpose tracing categories. Most tracing categories are PERL package
1558    names. So, for example, if you wanted to turn on tracing inside the B<Sprout>,
1559    B<ERDB>, and B<SproutLoad> packages, you would specify the categories
1560    
1561        ["Sprout", "SproutLoad", "ERDB"]
1562    
1563    This would cause trace messages in the specified three packages to appear in
1564    the output. There are two special tracing categories that are automatically
1565    handled by this method. In other words, if you used L</TSetup> you would need
1566    to include these categories manually, but if you use this method they are turned
1567    on automatically.
1568    
1569    =over 4
1570    
1571    =item SQL
1572    
1573    Traces SQL commands and activity.
1574    
1575    =item Tracer
1576    
1577    Traces error messages and call stacks.
1578    
1579    =back
1580    
1581    C<SQL> is only turned on if the C<-sql> option is specified in the command line.
1582    The trace level is specified using the C<-trace> command-line option. For example,
1583    the following command line for C<TransactFeatures> turns on SQL tracing and runs
1584    all tracing at level 3.
1585    
1586        TransactFeatures -trace=3 -sql register ../xacts IDs.tbl
1587    
1588    Standard tracing is output to the standard output and echoed to the file
1589    C<trace>I<$$>C<.log> in the FIG temporary directory, where I<$$> is the
1590    process ID. You can also specify the C<user> parameter to put a user ID
1591    instead of a process ID in the trace file name. So, for example
1592    
1593    The default trace level is 2. To get all messages, specify a trace level of 4.
1594    For a genome-by-genome update, use 3.
1595    
1596        TransactFeatures -trace=3 -sql -user=Bruce register ../xacts IDs.tbl
1597    
1598    would send the trace output to C<traceBruce.log> in the temporary directory.
1599    
1600    The I<options> parameter is a reference to a hash containing the command-line
1601    options, their default values, and an explanation of what they mean. Command-line
1602    options may be in the form of switches or keywords. In the case of a switch, the
1603    option value is 1 if it is specified and 0 if it is not specified. In the case
1604    of a keyword, the value is separated from the option name by an equal sign. You
1605    can see this last in the command-line example above.
1606    
1607    You can specify a different default trace level by setting C<$options->{trace}>
1608    prior to calling this method.
1609    
1610    An example at this point would help. Consider, for example, the command-line utility
1611    C<TransactFeatures>. It accepts a list of positional parameters plus the options
1612    C<safe>, C<noAlias>, C<start>, and C<tblFiles>. To start up this command, we execute
1613    the following code.
1614    
1615        my ($options, @parameters) = Tracer::StandardSetup(["DocUtils"],
1616                            { safe => [0, "use database transactions"],
1617                              noAlias => [0, "do not expect aliases in CHANGE transactions"],
1618                              start => [' ', "start with this genome"],
1619                              tblFiles => [0, "output TBL files containing the corrected IDs"] },
1620                            "<command> <transactionDirectory> <IDfile>",
1621                          @ARGV);
1622    
1623    
1624    The call to C<ParseCommand> specifies the default values for the options and
1625    stores the actual options in a hash that is returned as C<$options>. The
1626    positional parameters are returned in C<@parameters>.
1627    
1628    The following is a sample command line for C<TransactFeatures>.
1629    
1630        TransactFeatures -trace=2 -noAlias register ../xacts IDs.tbl
1631    
1632    Single and double hyphens are equivalent. So, you could also code the
1633    above command as
1634    
1635        TransactFeatures --trace=2 --noAlias register ../xacts IDs.tbl
1636    
1637    In this case, C<register>, C<../xacts>, and C<IDs.tbl> are the positional
1638    parameters, and would find themselves in I<@parameters> after executing the
1639    above code fragment. The tracing would be set to level 2, and the categories
1640    would be C<Tracer>, and <DocUtils>. C<Tracer> is standard,
1641    and C<DocUtils> was included because it came in within the first parameter
1642    to this method. The I<$options> hash would be
1643    
1644        { trace => 2, sql => 0, safe => 0,
1645          noAlias => 1, start => ' ', tblFiles => 0 }
1646    
1647    Use of C<StandardSetup> in this way provides a simple way of performing
1648    standard tracing setup and command-line parsing. Note that the caller is
1649    not even aware of the command-line switches C<-trace> and C<-sql>, which
1650    are used by this method to control the tracing. If additional tracing features
1651    need to be added in the future, they can be processed by this method without
1652    upsetting the command-line utilities.
1653    
1654    If the C<background> option is specified on the command line, then the
1655    standard and error outputs will be directed to files in the temporary
1656    directory, using the same suffix as the trace file. So, if the command
1657    line specified
1658    
1659        -user=Bruce -background
1660    
1661    then the trace output would go to C<traceBruce.log>, the standard output to
1662    C<outBruce.log>, and the error output to C<errBruce.log>. This is designed to
1663    simplify starting a command in the background.
1664    
1665    The user name is also used as the tracing key for L</Emergency Tracing>.
1666    Specifying a value of C<E> for the trace level causes emergency tracing to
1667    be used instead of custom tracing. If the user name is not specified,
1668    the tracing key is taken from the C<Tracing> environment variable. If there
1669    is no value for that variable, the tracing key will be computed from the active
1670    login ID.
1671    
1672    Since the default situation in StandardSetup is to trace to the standard
1673    output, errors that occur in command-line scripts will not generate
1674    RSS events. To force the events, use the C<warn> option.
1675    
1676        TransactFeatures -background -warn register ../xacts IDs.tbl
1677    
1678    Finally, if the special option C<-help> is specified, the option
1679    names will be traced at level 0 and the program will exit without processing.
1680    This provides a limited help capability. For example, if the user enters
1681    
1682        TransactFeatures -help
1683    
1684    he would see the following output.
1685    
1686        TransactFeatures [options] <command> <transactionDirectory> <IDfile>
1687            -trace    tracing level (default E)
1688            -sql      trace SQL commands
1689            -safe     use database transactions
1690            -noAlias  do not expect aliases in CHANGE transactions
1691            -start    start with this genome
1692            -tblFiles output TBL files containing the corrected IDs
1693    
1694    The caller has the option of modifying the tracing scheme by placing a value
1695    for C<trace> in the incoming options hash. The default value can be overridden,
1696    or the tracing to the standard output can be turned off by suffixing a minus
1697    sign to the trace level. So, for example,
1698    
1699        { trace => [0, "tracing level (default 0)"],
1700           ...
1701    
1702    would set the default trace level to 0 instead of E, while
1703    
1704        { trace => ["2-", "tracing level (default 2)"],
1705           ...
1706    
1707    would set the default to 2, but trace only to the log file, not to the
1708    standard output.
1709    
1710    The parameters to this method are as follows.
1711    
1712    =over 4
1713    
1714    =item categories
1715    
1716    Reference to a list of tracing category names. These should be names of
1717    packages whose internal workings will need to be debugged to get the
1718    command working.
1719    
1720    =item options
1721    
1722    Reference to a hash containing the legal options for the current command mapped
1723    to their default values and descriptions. The user can override the defaults
1724    by specifying the options as command-line switches prefixed by a hyphen.
1725    Tracing-related options may be added to this hash. If the C<-h> option is
1726    specified on the command line, the option descriptions will be used to
1727    explain the options. To turn off tracing to the standard output, add a
1728    minus sign to the value for C<trace> (see above).
1729    
1730    =item parmHelp
1731    
1732    A string that vaguely describes the positional parameters. This is used
1733    if the user specifies the C<-h> option.
1734    
1735    =item argv
1736    
1737    List of command line parameters, including the option switches, which must
1738    precede the positional parameters and be prefixed by a hyphen.
1739    
1740    =item RETURN
1741    
1742    Returns a list. The first element of the list is the reference to a hash that
1743    maps the command-line option switches to their values. These will either be the
1744    default values or overrides specified on the command line. The remaining
1745    elements of the list are the position parameters, in order.
1746    
1747    =back
1748    
1749    =cut
1750    
1751    sub StandardSetup {
1752        # Get the parameters.
1753        my ($categories, $options, $parmHelp, @argv) = @_;
1754        # Get the default tracing key.
1755        my $tkey = EmergencyKey();
1756        # Save the command line.
1757        $CommandLine = join(" ", $0, map { $_ =~ /\s/ ? "\"$_\"" : $_ } @argv);
1758        # Add the tracing options.
1759        if (! exists $options->{trace}) {
1760            $options->{trace} = ['2', "tracing level (E for emergency tracing)"];
1761        }
1762        $options->{sql} = [0, "turn on SQL tracing"];
1763        $options->{help} = [0, "display command-line options"];
1764        $options->{user} = [$tkey, "tracing key"];
1765        $options->{background} = [0, "spool standard and error output"];
1766        $options->{warn} = [0, "send errors to RSS feed"];
1767        # Create a parsing hash from the options hash. The parsing hash
1768        # contains the default values rather than the default value
1769        # and the description. While we're at it, we'll memorize the
1770        # length of the longest option name.
1771        my $longestName = 0;
1772      my %parseOptions = ();      my %parseOptions = ();
1773      for my $key (keys %{$options}) {      for my $key (keys %{$options}) {
1774          if (length $key > $longestName) {          if (length $key > $longestName) {
1775              $longestName = length $key;              $longestName = length $key;
1776          }          }
1777          $parseOptions{$key} = $options->{$key}->[0];          $parseOptions{$key} = $options->{$key}->[0];
1778        }
1779        # Parse the command line.
1780        my ($retOptions, @retParameters) = ParseCommand(\%parseOptions, @argv);
1781        # Get the logfile suffix.
1782        my $suffix = $retOptions->{user};
1783        # Check for background mode.
1784        if ($retOptions->{background}) {
1785            my $outFileName = "$FIG_Config::temp/out$suffix.log";
1786            my $errFileName = "$FIG_Config::temp/err$suffix.log";
1787            open STDOUT, ">$outFileName";
1788            open STDERR, ">$errFileName";
1789            # Check for phone support. If we have phone support and a phone number,
1790            # we want to turn it on.
1791            if ($ENV{PHONE} && defined($FIG_Config::phone)) {
1792                $retOptions->{phone} = $ENV{PHONE};
1793            }
1794        }
1795        # Now we want to set up tracing. First, we need to know if the user
1796        # wants emergency tracing.
1797        if ($retOptions->{trace} eq 'E') {
1798            ETracing($retOptions->{user});
1799        } else {
1800            # Here the tracing is controlled from the command line.
1801            my @cats = @{$categories};
1802            if ($retOptions->{sql}) {
1803                push @cats, "SQL";
1804            }
1805            if ($retOptions->{warn}) {
1806                push @cats, "Feed";
1807            }
1808            # Add the default categories.
1809            push @cats, "Tracer";
1810            # Next, we create the category string by joining the categories.
1811            my $cats = join(" ", @cats);
1812            # Check to determine whether or not the caller wants to turn off tracing
1813            # to the standard output.
1814            my $traceLevel = $retOptions->{trace};
1815            my $textOKFlag = 1;
1816            if ($traceLevel =~ /^(.)-/) {
1817                $traceLevel = $1;
1818                $textOKFlag = 0;
1819            }
1820            # Now we set up the trace mode.
1821            my $traceMode;
1822            # Verify that we can open a file in the FIG temporary directory.
1823            my $traceFileName = "$FIG_Config::temp/trace$suffix.log";
1824            if (open TESTTRACE, ">$traceFileName") {
1825                # Here we can trace to a file.
1826                $traceMode = ">$traceFileName";
1827                if ($textOKFlag) {
1828                    # Echo to standard output if the text-OK flag is set.
1829                    $traceMode = "+$traceMode";
1830                }
1831                # Close the test file.
1832                close TESTTRACE;
1833            } else {
1834                # Here we can't trace to a file. Complain about this.
1835                warn "Could not open trace file $traceFileName: $!\n";
1836                # We trace to the standard output if it's
1837                # okay, and the error log otherwise.
1838                if ($textOKFlag) {
1839                    $traceMode = "TEXT";
1840                } else {
1841                    $traceMode = "WARN";
1842                }
1843            }
1844            # Now set up the tracing.
1845            TSetup("$traceLevel $cats", $traceMode);
1846        }
1847        # Check for the "help" option. If it is specified, dump the command-line
1848        # options and exit the program.
1849        if ($retOptions->{help}) {
1850            $0 =~ m#[/\\](\w+)(\.pl)?$#i;
1851            print "$1 [options] $parmHelp\n";
1852            for my $key (sort keys %{$options}) {
1853                my $name = Pad($key, $longestName, 0, ' ');
1854                my $desc = $options->{$key}->[1];
1855                if ($options->{$key}->[0]) {
1856                    $desc .= " (default " . $options->{$key}->[0] . ")";
1857                }
1858                print "  $name $desc\n";
1859            }
1860            exit(0);
1861        }
1862        # Trace the options, if applicable.
1863        if (T(3)) {
1864            my @parms = grep { $retOptions->{$_} } keys %{$retOptions};
1865            Trace("Selected options: " . join(", ", sort @parms) . ".");
1866        }
1867        # Return the parsed parameters.
1868        return ($retOptions, @retParameters);
1869    }
1870    
1871    =head3 ReadOptions
1872    
1873        my %options = Tracer::ReadOptions($fileName);
1874    
1875    Read a set of options from a file. Each option is encoded in a line of text that has the
1876    format
1877    
1878    I<optionName>C<=>I<optionValue>C<; >I<comment>
1879    
1880    The option name must consist entirely of letters, digits, and the punctuation characters
1881    C<.> and C<_>, and is case sensitive. Blank lines and lines in which the first nonblank
1882    character is a semi-colon will be ignored. The return hash will map each option name to
1883    the corresponding option value.
1884    
1885    =over 4
1886    
1887    =item fileName
1888    
1889    Name of the file containing the option data.
1890    
1891    =item RETURN
1892    
1893    Returns a hash mapping the option names specified in the file to their corresponding option
1894    value.
1895    
1896    =back
1897    
1898    =cut
1899    
1900    sub ReadOptions {
1901        # Get the parameters.
1902        my ($fileName) = @_;
1903        # Open the file.
1904        (open CONFIGFILE, "<$fileName") || Confess("Could not open option file $fileName.");
1905        # Count the number of records read.
1906        my ($records, $comments) = 0;
1907        # Create the return hash.
1908        my %retVal = ();
1909        # Loop through the file, accumulating key-value pairs.
1910        while (my $line = <CONFIGFILE>) {
1911            # Denote we've read a line.
1912            $records++;
1913            # Determine the line type.
1914            if ($line =~ /^\s*[\n\r]/) {
1915                # A blank line is a comment.
1916                $comments++;
1917            } elsif ($line =~ /^\s*([A-Za-z0-9_\.]+)=([^;]*);/) {
1918                # Here we have an option assignment.
1919                retVal{$1} = $2;
1920            } elsif ($line =~ /^\s*;/) {
1921                # Here we have a text comment.
1922                $comments++;
1923            } else {
1924                # Here we have an invalid line.
1925                Trace("Invalid option statement in record $records.") if T(0);
1926            }
1927        }
1928        # Return the hash created.
1929        return %retVal;
1930    }
1931    
1932    =head3 GetOptions
1933    
1934        Tracer::GetOptions(\%defaults, \%options);
1935    
1936    Merge a specified set of options into a table of defaults. This method takes two hash references
1937    as input and uses the data from the second to update the first. If the second does not exist,
1938    there will be no effect. An error will be thrown if one of the entries in the second hash does not
1939    exist in the first.
1940    
1941    Consider the following example.
1942    
1943        my $optionTable = GetOptions({ dbType => 'mySQL', trace => 0 }, $options);
1944    
1945    In this example, the variable B<$options> is expected to contain at most two options-- B<dbType> and
1946    B<trace>. The default database type is C<mySQL> and the default trace level is C<0>. If the value of
1947    B<$options> is C<< {dbType => 'Oracle'} >>, then the database type will be changed to C<Oracle> and
1948    the trace level will remain at 0. If B<$options> is undefined, then the database type and trace level
1949    will remain C<mySQL> and C<0>. If, on the other hand, B<$options> is defined as
1950    
1951        {databaseType => 'Oracle'}
1952    
1953    an error will occur because the B<databaseType> option does not exist.
1954    
1955    =over 4
1956    
1957    =item defaults
1958    
1959    Table of default option values.
1960    
1961    =item options
1962    
1963    Table of overrides, if any.
1964    
1965    =item RETURN
1966    
1967    Returns a reference to the default table passed in as the first parameter.
1968    
1969    =back
1970    
1971    =cut
1972    
1973    sub GetOptions {
1974        # Get the parameters.
1975        my ($defaults, $options) = @_;
1976        # Check for overrides.
1977        if ($options) {
1978            # Loop through the overrides.
1979            while (my ($option, $setting) = each %{$options}) {
1980                # Insure this override exists.
1981                if (!exists $defaults->{$option}) {
1982                    croak "Unrecognized option $option encountered.";
1983                } else {
1984                    # Apply the override.
1985                    $defaults->{$option} = $setting;
1986                }
1987            }
1988        }
1989        # Return the merged table.
1990        return $defaults;
1991    }
1992    
1993    =head3 MergeOptions
1994    
1995        Tracer::MergeOptions(\%table, \%defaults);
1996    
1997    Merge default values into a hash table. This method looks at the key-value pairs in the
1998    second (default) hash, and if a matching key is not found in the first hash, the default
1999    pair is copied in. The process is similar to L</GetOptions>, but there is no error-
2000    checking and no return value.
2001    
2002    =over 4
2003    
2004    =item table
2005    
2006    Hash table to be updated with the default values.
2007    
2008    =item defaults
2009    
2010    Default values to be merged into the first hash table if they are not already present.
2011    
2012    =back
2013    
2014    =cut
2015    
2016    sub MergeOptions {
2017        # Get the parameters.
2018        my ($table, $defaults) = @_;
2019        # Loop through the defaults.
2020        while (my ($key, $value) = each %{$defaults}) {
2021            if (!exists $table->{$key}) {
2022                $table->{$key} = $value;
2023            }
2024        }
2025      }      }
2026      # Parse the command line.  
2027      my ($retOptions, @retParameters) = ParseCommand(\%parseOptions, @argv);  =head3 ParseCommand
2028      # Get the logfile suffix.  
2029      my $suffix = $retOptions->{user};      my ($options, @arguments) = Tracer::ParseCommand(\%optionTable, @inputList);
2030      # Check for background mode.  
2031      if ($retOptions->{background}) {  Parse a command line consisting of a list of parameters. The initial parameters may be option
2032          my $outFileName = "$FIG_Config::temp/out$suffix.log";  specifiers of the form C<->I<option> or C<->I<option>C<=>I<value>. The options are stripped
2033          my $errFileName = "$FIG_Config::temp/err$suffix.log";  off and merged into a table of default options. The remainder of the command line is
2034          open STDOUT, ">$outFileName";  returned as a list of positional arguments. For example, consider the following invocation.
2035          open STDERR, ">$errFileName";  
2036        my ($options, @arguments) = ParseCommand({ errors => 0, logFile => 'trace.log'}, @words);
2037    
2038    In this case, the list @words will be treated as a command line and there are two options available,
2039    B<errors> and B<logFile>. If @words has the following format
2040    
2041        -logFile=error.log apple orange rutabaga
2042    
2043    then at the end of the invocation, C<$options> will be
2044    
2045        { errors => 0, logFile => 'error.log' }
2046    
2047    and C<@arguments> will contain
2048    
2049        apple orange rutabaga
2050    
2051    The parser allows for some escape sequences. See L</UnEscape> for a description. There is no
2052    support for quote characters. Options can be specified with single or double hyphens.
2053    
2054    =over 4
2055    
2056    =item optionTable
2057    
2058    Table of default options.
2059    
2060    =item inputList
2061    
2062    List of words on the command line.
2063    
2064    =item RETURN
2065    
2066    Returns a reference to the option table and a list of the positional arguments.
2067    
2068    =back
2069    
2070    =cut
2071    
2072    sub ParseCommand {
2073        # Get the parameters.
2074        my ($optionTable, @inputList) = @_;
2075        # Process any options in the input list.
2076        my %overrides = ();
2077        while ((@inputList > 0) && ($inputList[0] =~ /^--?/)) {
2078            # Get the current option.
2079            my $arg = shift @inputList;
2080            # Pull out the option name.
2081            $arg =~ /^--?([^=]*)/g;
2082            my $name = $1;
2083            # Check for an option value.
2084            if ($arg =~ /\G=(.*)$/g) {
2085                # Here we have a value for the option.
2086                $overrides{$name} = UnEscape($1);
2087            } else {
2088                # Here there is no value, so we use 1.
2089                $overrides{$name} = 1;
2090      }      }
     # Now we want to set up tracing. First, we need to know if SQL is to  
     # be traced.  
     my @cats = @{$categories};  
     if ($retOptions->{sql}) {  
         push @cats, "SQL";  
2091      }      }
2092      # Add the default categories.      # Merge the options into the defaults.
2093      push @cats, "Tracer", "FIG";      GetOptions($optionTable, \%overrides);
2094      # Next, we create the category string by joining the categories.      # Translate the remaining parameters.
2095      my $cats = join(" ", @cats);      my @retVal = ();
2096      # Check to determine whether or not the caller wants to turn off tracing      for my $inputParm (@inputList) {
2097      # to the standard output.          push @retVal, UnEscape($inputParm);
     my $traceLevel = $retOptions->{trace};  
     my $textOKFlag = 1;  
     if ($traceLevel =~ /^(.)-/) {  
         $traceLevel = $1;  
         $textOKFlag = 0;  
2098      }      }
2099      # Now we set up the trace mode.      # Return the results.
2100      my $traceMode;      return ($optionTable, @retVal);
     # Verify that we can open a file in the FIG temporary directory.  
     my $traceFileName = "$FIG_Config::temp/trace$suffix.log";  
     if (open TESTTRACE, ">$traceFileName") {  
         # Here we can trace to a file.  
         $traceMode = ">$traceFileName";  
         if ($textOKFlag) {  
             # Echo to standard output if the text-OK flag is set.  
             $traceMode = "+$traceMode";  
2101          }          }
2102          # Close the test file.  
2103          close TESTTRACE;  
2104      } else {  =head2 File Utility Methods
2105          # Here we can't trace to a file. We trace to the standard output if it's  
2106          # okay, and the error log otherwise.  =head3 GetFile
2107          if ($textOKFlag) {  
2108              $traceMode = "TEXT";      my @fileContents = Tracer::GetFile($fileName);
2109    
2110        or
2111    
2112        my $fileContents = Tracer::GetFile($fileName);
2113    
2114    Return the entire contents of a file. In list context, line-ends are removed and
2115    each line is a list element. In scalar context, line-ends are replaced by C<\n>.
2116    
2117    =over 4
2118    
2119    =item fileName
2120    
2121    Name of the file to read.
2122    
2123    =item RETURN
2124    
2125    In a list context, returns the entire file as a list with the line terminators removed.
2126    In a scalar context, returns the entire file as a string. If an error occurs opening
2127    the file, an empty list will be returned.
2128    
2129    =back
2130    
2131    =cut
2132    
2133    sub GetFile {
2134        # Get the parameters.
2135        my ($fileName) = @_;
2136        # Declare the return variable.
2137        my @retVal = ();
2138        # Open the file for input.
2139        my $handle = Open(undef, "<$fileName");
2140        # Read the whole file into the return variable, stripping off any terminator
2141        # characters.
2142        my $lineCount = 0;
2143        while (my $line = <$handle>) {
2144            $lineCount++;
2145            $line = Strip($line);
2146            push @retVal, $line;
2147        }
2148        # Close it.
2149        close $handle;
2150        my $actualLines = @retVal;
2151        Trace("$actualLines lines read from file $fileName.") if T(File => 2);
2152        # Return the file's contents in the desired format.
2153        if (wantarray) {
2154            return @retVal;
2155          } else {          } else {
2156              $traceMode = "WARN";          return join "\n", @retVal;
2157          }          }
2158      }      }
2159      # Now set up the tracing.  
2160      TSetup("$traceLevel $cats", $traceMode);  =head3 PutFile
2161      # Check for the "h" option. If it is specified, dump the command-line  
2162      # options and exit the program.      Tracer::PutFile($fileName, \@lines);
2163      if ($retOptions->{h}) {  
2164          $0 =~ m#[/\\](\w+)(\.pl)?$#i;  Write out a file from a list of lines of text.
2165          Trace("$1 [options] $parmHelp") if T(0);  
2166          for my $key (sort keys %{$options}) {  =over 4
2167              my $name = Pad($key, $longestName, 0, ' ');  
2168              my $desc = $options->{$key}->[1];  =item fileName
2169              if ($options->{$key}->[0]) {  
2170                  $desc .= " (default " . $options->{$key}->[0] . ")";  Name of the output file.
2171    
2172    =item lines
2173    
2174    Reference to a list of text lines. The lines will be written to the file in order, with trailing
2175    new-line characters. Alternatively, may be a string, in which case the string will be written without
2176    modification.
2177    
2178    =back
2179    
2180    =cut
2181    
2182    sub PutFile {
2183        # Get the parameters.
2184        my ($fileName, $lines) = @_;
2185        # Open the output file.
2186        my $handle = Open(undef, ">$fileName");
2187        # Count the lines written.
2188        if (ref $lines ne 'ARRAY') {
2189            # Here we have a scalar, so we write it raw.
2190            print $handle $lines;
2191            Trace("Scalar put to file $fileName.") if T(File => 3);
2192        } else {
2193            # Write the lines one at a time.
2194            my $count = 0;
2195            for my $line (@{$lines}) {
2196                print $handle "$line\n";
2197                $count++;
2198              }              }
2199              Trace("  $name $desc") if T(0);          Trace("$count lines put to file $fileName.") if T(File => 3);
2200          }          }
2201          exit(0);      # Close the output file.
2202        close $handle;
2203      }      }
2204      # Return the parsed parameters.  
2205      return ($retOptions, @retParameters);  =head3 ParseRecord
2206    
2207        my @fields = Tracer::ParseRecord($line);
2208    
2209    Parse a tab-delimited data line. The data line is split into field values. Embedded tab
2210    and new-line characters in the data line must be represented as C<\t> and C<\n>, respectively.
2211    These will automatically be converted.
2212    
2213    =over 4
2214    
2215    =item line
2216    
2217    Line of data containing the tab-delimited fields.
2218    
2219    =item RETURN
2220    
2221    Returns a list of the fields found in the data line.
2222    
2223    =back
2224    
2225    =cut
2226    
2227    sub ParseRecord {
2228        # Get the parameter.
2229        my ($line) = @_;
2230        # Remove the trailing new-line, if any.
2231        chomp $line;
2232        # Split the line read into pieces using the tab character.
2233        my @retVal = split /\t/, $line;
2234        # Trim and fix the escapes in each piece.
2235        for my $value (@retVal) {
2236            # Trim leading whitespace.
2237            $value =~ s/^\s+//;
2238            # Trim trailing whitespace.
2239            $value =~ s/\s+$//;
2240            # Delete the carriage returns.
2241            $value =~ s/\r//g;
2242            # Convert the escapes into their real values.
2243            $value =~ s/\\t/"\t"/ge;
2244            $value =~ s/\\n/"\n"/ge;
2245        }
2246        # Return the result.
2247        return @retVal;
2248  }  }
2249    
2250  =head3 Setups  =head3 Merge
2251    
2252  C<< my $count = Tracer::Setups(); >>      my @mergedList = Tracer::Merge(@inputList);
2253    
2254  Return the number of times L</TSetup> has been called.  Sort a list of strings and remove duplicates.
2255    
2256  This method allows for the creation of conditional tracing setups where, for example, we  =over 4
2257  may want to set up tracing if nobody else has done it before us.  
2258    =item inputList
2259    
2260    List of scalars to sort and merge.
2261    
2262    =item RETURN
2263    
2264    Returns a list containing the same elements sorted in ascending order with duplicates
2265    removed.
2266    
2267    =back
2268    
2269  =cut  =cut
2270    
2271  sub Setups {  sub Merge {
2272      return $SetupCount;      # Get the input list in sort order.
2273        my @inputList = sort @_;
2274        # Only proceed if the list has at least two elements.
2275        if (@inputList > 1) {
2276            # Now we want to move through the list splicing out duplicates.
2277            my $i = 0;
2278            while ($i < @inputList) {
2279                # Get the current entry.
2280                my $thisEntry = $inputList[$i];
2281                # Find out how many elements duplicate the current entry.
2282                my $j = $i + 1;
2283                my $dup1 = $i + 1;
2284                while ($j < @inputList && $inputList[$j] eq $thisEntry) { $j++; };
2285                # If the number is nonzero, splice out the duplicates found.
2286                if ($j > $dup1) {
2287                    splice @inputList, $dup1, $j - $dup1;
2288                }
2289                # Now the element at position $dup1 is different from the element before it
2290                # at position $i. We push $i forward one position and start again.
2291                $i++;
2292            }
2293        }
2294        # Return the merged list.
2295        return @inputList;
2296  }  }
2297    
2298  =head3 Open  =head3 Open
2299    
2300  C<< my $handle = Open($fileHandle, $fileSpec, $message); >>      my $handle = Open($fileHandle, $fileSpec, $message);
2301    
2302  Open a file.  Open a file.
2303    
# Line 603  Line 2385 
2385      }      }
2386      # Return the file handle.      # Return the file handle.
2387      return $fileHandle;      return $fileHandle;
2388  }  }
   
 =head3 FindNamePart  
   
 C<< my ($fileName, $start, $len) = Tracer::FindNamePart($fileSpec); >>  
   
 Extract the portion of a file specification that contains the file name.  
   
 A file specification is the string passed to an C<open> call. It specifies the file  
 mode and name. In a truly complex situation, it can specify a pipe sequence. This  
 method assumes that the file name is whatever follows the first angle bracket  
 sequence.  So, for example, in the following strings the file name is  
 C</usr/fig/myfile.txt>.  
   
     >>/usr/fig/myfile.txt  
     </usr/fig/myfile.txt  
     | sort -u > /usr/fig/myfile.txt  
   
 If the method cannot find a file name using its normal methods, it will return the  
 whole incoming string.  
   
 =over 4  
   
 =item fileSpec  
   
 File specification string from which the file name is to be extracted.  
   
 =item RETURN  
   
 Returns a three-element list. The first element contains the file name portion of  
 the specified string, or the whole string if a file name cannot be found via normal  
 methods. The second element contains the start position of the file name portion and  
 the third element contains the length.  
   
 =back  
   
 =cut  
 #: Return Type $;  
 sub FindNamePart {  
     # Get the parameters.  
     my ($fileSpec) = @_;  
     # Default to the whole input string.  
     my ($retVal, $pos, $len) = ($fileSpec, 0, length $fileSpec);  
     # Parse out the file name if we can.  
     if ($fileSpec =~ m/(<|>>?)(.+?)(\s*)$/) {  
         $retVal = $2;  
         $len = length $retVal;  
         $pos = (length $fileSpec) - (length $3) - $len;  
     }  
     # Return the result.  
     return ($retVal, $pos, $len);  
 }  
   
 =head3 OpenDir  
   
 C<< my @files = OpenDir($dirName, $filtered, $flag); >>  
   
 Open a directory and return all the file names. This function essentially performs  
 the functions of an C<opendir> and C<readdir>. If the I<$filtered> parameter is  
 set to TRUE, all filenames beginning with a period (C<.>), dollar sign (C<$>),  
 or pound sign (C<#>) and all filenames ending with a tilde C<~>) will be  
 filtered out of the return list. If the directory does not open and I<$flag> is not  
 set, an exception is thrown. So, for example,  
   
     my @files = OpenDir("/Volumes/fig/contigs", 1);  
   
 is effectively the same as  
   
     opendir(TMP, "/Volumes/fig/contigs") || Confess("Could not open /Volumes/fig/contigs.");  
     my @files = grep { $_ !~ /^[\.\$\#]/ && $_ !~ /~$/ } readdir(TMP);  
   
 Similarly, the following code  
   
     my @files = grep { $_ =~ /^\d/ } OpenDir("/Volumes/fig/orgs", 0, 1);  
   
 Returns the names of all files in C</Volumes/fig/orgs> that begin with digits and  
 automatically returns an empty list if the directory fails to open.  
   
 =over 4  
   
 =item dirName  
   
 Name of the directory to open.  
   
 =item filtered  
   
 TRUE if files whose names begin with a period (C<.>) should be automatically removed  
 from the list, else FALSE.  
   
 =item flag  
   
 TRUE if a failure to open is okay, else FALSE  
   
 =back  
   
 =cut  
 #: Return Type @;  
 sub OpenDir {  
     # Get the parameters.  
     my ($dirName, $filtered, $flag) = @_;  
     # Declare the return variable.  
     my @retVal = ();  
     # Open the directory.  
     if (opendir(my $dirHandle, $dirName)) {  
         # The directory opened successfully. Get the appropriate list according to the  
         # strictures of the filter parameter.  
         if ($filtered) {  
             @retVal = grep { $_ !~ /^[\.\$\#]/ && $_ !~ /~$/ } readdir $dirHandle;  
         } else {  
             @retVal = readdir $dirHandle;  
         }  
     } elsif (! $flag) {  
         # Here the directory would not open and it's considered an error.  
         Confess("Could not open directory $dirName.");  
     }  
     # Return the result.  
     return @retVal;  
 }  
   
 =head3 SetLevel  
   
 C<< Tracer::SetLevel($newLevel); >>  
   
 Modify the trace level. A higher trace level will cause more messages to appear.  
   
 =over 4  
   
 =item newLevel  
   
 Proposed new trace level.  
   
 =back  
   
 =cut  
   
 sub SetLevel {  
     $TraceLevel = $_[0];  
 }  
   
 =head3 Now  
   
 C<< my $string = Tracer::Now(); >>  
   
 Return a displayable time stamp containing the local time.  
   
 =cut  
   
 sub Now {  
     my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime(time);  
     my $retVal = _p2($mon+1) . "/" . _p2($mday) . "/" . ($year + 1900) . " " .  
                  _p2($hour) . ":" . _p2($min) . ":" . _p2($sec);  
     return $retVal;  
 }  
   
 # Pad a number to 2 digits.  
 sub _p2 {  
     my ($value) = @_;  
     $value = "0$value" if ($value < 10);  
     return $value;  
 }  
   
 =head3 LogErrors  
   
 C<< Tracer::LogErrors($fileName); >>  
   
 Route the standard error output to a log file.  
   
 =over 4  
   
 =item fileName  
   
 Name of the file to receive the error output.  
   
 =back  
   
 =cut  
   
 sub LogErrors {  
     # Get the file name.  
     my ($fileName) = @_;  
     # Open the file as the standard error output.  
     open STDERR, '>', $fileName;  
 }  
   
 =head3 ReadOptions  
   
 C<< my %options = Tracer::ReadOptions($fileName); >>  
   
 Read a set of options from a file. Each option is encoded in a line of text that has the  
 format  
   
 I<optionName>C<=>I<optionValue>C<; >I<comment>  
   
 The option name must consist entirely of letters, digits, and the punctuation characters  
 C<.> and C<_>, and is case sensitive. Blank lines and lines in which the first nonblank  
 character is a semi-colon will be ignored. The return hash will map each option name to  
 the corresponding option value.  
   
 =over 4  
   
 =item fileName  
   
 Name of the file containing the option data.  
   
 =item RETURN  
   
 Returns a hash mapping the option names specified in the file to their corresponding option  
 value.  
   
 =back  
   
 =cut  
   
 sub ReadOptions {  
     # Get the parameters.  
     my ($fileName) = @_;  
     # Open the file.  
     (open CONFIGFILE, "<$fileName") || Confess("Could not open option file $fileName.");  
     # Count the number of records read.  
     my ($records, $comments) = 0;  
     # Create the return hash.  
     my %retVal = ();  
     # Loop through the file, accumulating key-value pairs.  
     while (my $line = <CONFIGFILE>) {  
         # Denote we've read a line.  
         $records++;  
         # Determine the line type.  
         if ($line =~ /^\s*[\n\r]/) {  
             # A blank line is a comment.  
             $comments++;  
         } elsif ($line =~ /^\s*([A-Za-z0-9_\.]+)=([^;]*);/) {  
             # Here we have an option assignment.  
             retVal{$1} = $2;  
         } elsif ($line =~ /^\s*;/) {  
             # Here we have a text comment.  
             $comments++;  
         } else {  
             # Here we have an invalid line.  
             Trace("Invalid option statement in record $records.") if T(0);  
         }  
     }  
     # Return the hash created.  
     return %retVal;  
 }  
   
 =head3 GetOptions  
   
 C<< Tracer::GetOptions(\%defaults, \%options); >>  
   
 Merge a specified set of options into a table of defaults. This method takes two hash references  
 as input and uses the data from the second to update the first. If the second does not exist,  
 there will be no effect. An error will be thrown if one of the entries in the second hash does not  
 exist in the first.  
   
 Consider the following example.  
2389    
2390  C<< my $optionTable = GetOptions({ dbType => 'mySQL', trace => 0 }, $options); >>  =head3 FindNamePart
2391    
2392  In this example, the variable B<$options> is expected to contain at most two options-- B<dbType> and      my ($fileName, $start, $len) = Tracer::FindNamePart($fileSpec);
 B<trace>. The default database type is C<mySQL> and the default trace level is C<0>. If the value of  
 B<$options> is C<< {dbType => 'Oracle'} >>, then the database type will be changed to C<Oracle> and  
 the trace level will remain at 0. If B<$options> is undefined, then the database type and trace level  
 will remain C<mySQL> and C<0>. If, on the other hand, B<$options> is defined as  
2393    
2394  C<< {databaseType => 'Oracle'} >>  Extract the portion of a file specification that contains the file name.
2395    
2396  an error will occur because the B<databaseType> option does not exist.  A file specification is the string passed to an C<open> call. It specifies the file
2397    mode and name. In a truly complex situation, it can specify a pipe sequence. This
2398    method assumes that the file name is whatever follows the first angle bracket
2399    sequence.  So, for example, in the following strings the file name is
2400    C</usr/fig/myfile.txt>.
2401    
2402  =over 4      >>/usr/fig/myfile.txt
2403        </usr/fig/myfile.txt
2404        | sort -u > /usr/fig/myfile.txt
2405    
2406  =item defaults  If the method cannot find a file name using its normal methods, it will return the
2407    whole incoming string.
2408    
2409  Table of default option values.  =over 4
2410    
2411  =item options  =item fileSpec
2412    
2413  Table of overrides, if any.  File specification string from which the file name is to be extracted.
2414    
2415  =item RETURN  =item RETURN
2416    
2417  Returns a reference to the default table passed in as the first parameter.  Returns a three-element list. The first element contains the file name portion of
2418    the specified string, or the whole string if a file name cannot be found via normal
2419    methods. The second element contains the start position of the file name portion and
2420    the third element contains the length.
2421    
2422  =back  =back
2423    
2424  =cut  =cut
2425    #: Return Type $;
2426  sub GetOptions {  sub FindNamePart {
2427      # Get the parameters.      # Get the parameters.
2428      my ($defaults, $options) = @_;      my ($fileSpec) = @_;
2429      # Check for overrides.      # Default to the whole input string.
2430      if ($options) {      my ($retVal, $pos, $len) = ($fileSpec, 0, length $fileSpec);
2431          # Loop through the overrides.      # Parse out the file name if we can.
2432          while (my ($option, $setting) = each %{$options}) {      if ($fileSpec =~ m/(<|>>?)(.+?)(\s*)$/) {
2433              # Insure this override exists.          $retVal = $2;
2434              if (!exists $defaults->{$option}) {          $len = length $retVal;
2435                  croak "Unrecognized option $option encountered.";          $pos = (length $fileSpec) - (length $3) - $len;
             } else {  
                 # Apply the override.  
                 $defaults->{$option} = $setting;  
             }  
         }  
2436      }      }
2437      # Return the merged table.      # Return the result.
2438      return $defaults;      return ($retVal, $pos, $len);
2439  }  }
2440    
2441  =head3 MergeOptions  =head3 OpenDir
2442    
2443  C<< Tracer::MergeOptions(\%table, \%defaults); >>      my @files = OpenDir($dirName, $filtered, $flag);
2444    
2445  Merge default values into a hash table. This method looks at the key-value pairs in the  Open a directory and return all the file names. This function essentially performs
2446  second (default) hash, and if a matching key is not found in the first hash, the default  the functions of an C<opendir> and C<readdir>. If the I<$filtered> parameter is
2447  pair is copied in. The process is similar to L</GetOptions>, but there is no error-  set to TRUE, all filenames beginning with a period (C<.>), dollar sign (C<$>),
2448  checking and no return value.  or pound sign (C<#>) and all filenames ending with a tilde C<~>) will be
2449    filtered out of the return list. If the directory does not open and I<$flag> is not
2450    set, an exception is thrown. So, for example,
2451    
2452        my @files = OpenDir("/Volumes/fig/contigs", 1);
2453    
2454    is effectively the same as
2455    
2456        opendir(TMP, "/Volumes/fig/contigs") || Confess("Could not open /Volumes/fig/contigs.");
2457        my @files = grep { $_ !~ /^[\.\$\#]/ && $_ !~ /~$/ } readdir(TMP);
2458    
2459    Similarly, the following code
2460    
2461        my @files = grep { $_ =~ /^\d/ } OpenDir("/Volumes/fig/orgs", 0, 1);
2462    
2463    Returns the names of all files in C</Volumes/fig/orgs> that begin with digits and
2464    automatically returns an empty list if the directory fails to open.
2465    
2466  =over 4  =over 4
2467    
2468  =item table  =item dirName
2469    
2470  Hash table to be updated with the default values.  Name of the directory to open.
2471    
2472  =item defaults  =item filtered
2473    
2474  Default values to be merged into the first hash table if they are not already present.  TRUE if files whose names begin with a period (C<.>) should be automatically removed
2475    from the list, else FALSE.
2476    
2477    =item flag
2478    
2479    TRUE if a failure to open is okay, else FALSE
2480    
2481  =back  =back
2482    
2483  =cut  =cut
2484    #: Return Type @;
2485  sub MergeOptions {  sub OpenDir {
2486      # Get the parameters.      # Get the parameters.
2487      my ($table, $defaults) = @_;      my ($dirName, $filtered, $flag) = @_;
2488      # Loop through the defaults.      # Declare the return variable.
2489      while (my ($key, $value) = each %{$defaults}) {      my @retVal = ();
2490          if (!exists $table->{$key}) {      # Open the directory.
2491              $table->{$key} = $value;      if (opendir(my $dirHandle, $dirName)) {
2492            # The directory opened successfully. Get the appropriate list according to the
2493            # strictures of the filter parameter.
2494            if ($filtered) {
2495                @retVal = grep { $_ !~ /^[\.\$\#]/ && $_ !~ /~$/ } readdir $dirHandle;
2496            } else {
2497                @retVal = readdir $dirHandle;
2498          }          }
2499        } elsif (! $flag) {
2500            # Here the directory would not open and it's considered an error.
2501            Confess("Could not open directory $dirName.");
2502      }      }
2503        # Return the result.
2504        return @retVal;
2505  }  }
2506    
 =head3 Trace  
2507    
2508  C<< Trace($message); >>  =head3 Insure
2509    
2510  Write a trace message to the target location specified in L</TSetup>. If there has not been      Insure($dirName, $chmod);
2511  any prior call to B<TSetup>.  
2512    Insure a directory is present.
2513    
2514  =over 4  =over 4
2515    
2516  =item message  =item dirName
2517    
2518  Message to write.  Name of the directory to check. If it does not exist, it will be created.
2519    
2520    =item chmod (optional)
2521    
2522    Security privileges to be given to the directory if it is created.
2523    
2524  =back  =back
2525    
2526  =cut  =cut
2527    
2528  sub Trace {  sub Insure {
2529      # Get the parameters.      my ($dirName, $chmod) = @_;
2530      my ($message) = @_;      if (! -d $dirName) {
2531      # Get the timestamp.          Trace("Creating $dirName directory.") if T(2);
2532      my $timeStamp = Now();          eval {
2533      # Format the message. Note we strip off any line terminators at the end.              mkpath $dirName;
2534      my $formatted = "$timeStamp <$LastCategory>: " . Strip($message);              # If we have permissions specified, set them here.
2535      # Process according to the destination.              if (defined($chmod)) {
2536      if ($Destination eq "TEXT") {                  chmod $chmod, $dirName;
2537          # Write the message to the standard output.              }
2538          print "$formatted\n";          };
2539      } elsif ($Destination eq "ERROR") {          if ($@) {
2540          # Write the message to the error output.              Confess("Error creating $dirName: $@");
         print STDERR "$formatted\n";  
     } elsif ($Destination eq "QUEUE") {  
         # Push the message into the queue.  
         push @Queue, "$formatted";  
     } elsif ($Destination eq "HTML") {  
         # Convert the message to HTML and write it to the standard output.  
         my $escapedMessage = CGI::escapeHTML($message);  
         print "<p>$formatted</p>\n";  
     } elsif ($Destination eq "WARN") {  
        # Emit the message as a warning.  
        warn $message;  
     } elsif ($Destination =~ m/^>>/) {  
         # Write the trace message to an output file.  
         (open TRACING, $Destination) || die "Tracing open for \"$Destination\" failed: $!";  
         print TRACING "$formatted\n";  
         close TRACING;  
         # If the Tee flag is on, echo it to the standard output.  
         if ($TeeFlag) {  
             print "$formatted\n";  
2541          }          }
2542      }      }
2543  }  }
2544    
2545  =head3 T  =head3 ChDir
   
 C<< my $switch = T($category, $traceLevel); >>  
   
     or  
2546    
2547  C<< my $switch = T($traceLevel); >>      ChDir($dirName);
2548    
2549  Return TRUE if the trace level is at or above a specified value and the specified category  Change to the specified directory.
 is active, else FALSE. If no category is specified, the caller's package name is used.  
2550    
2551  =over 4  =over 4
2552    
2553  =item category  =item dirName
   
 Category to which the message belongs. If not specified, the caller's package name is  
 used.  
   
 =item traceLevel  
   
 Relevant tracing level.  
   
 =item RETURN  
2554    
2555  TRUE if a message at the specified trace level would appear in the trace, else FALSE.  Name of the directory to which we want to change.
2556    
2557  =back  =back
2558    
2559  =cut  =cut
2560    
2561  sub T {  sub ChDir {
2562      # Declare the return variable.      my ($dirName) = @_;
2563      my $retVal = 0;      if (! -d $dirName) {
2564      # Only proceed if tracing is turned on.          Confess("Cannot change to directory $dirName: no such directory.");
     if ($Destination ne "NONE") {  
         # Get the parameters.  
         my ($category, $traceLevel) = @_;  
         if (!defined $traceLevel) {  
             # Here we have no category, so we need to get the calling package.  
             # The calling package is normally the first parameter. If it is  
             # omitted, the first parameter will be the tracelevel. So, the  
             # first thing we do is shift the so-called category into the  
             # $traceLevel variable where it belongs.  
             $traceLevel = $category;  
             my ($package, $fileName, $line) = caller;  
             # If there is no calling package, we default to "main".  
             if (!$package) {  
                 $category = "main";  
2565              } else {              } else {
2566                  $category = $package;          Trace("Changing to directory $dirName.") if T(File => 4);
2567              }          my $okFlag = chdir $dirName;
2568          }          if (! $okFlag) {
2569          # Save the category name.              Confess("Error switching to directory $dirName.");
         $LastCategory = $category;  
         # Convert it to lower case before we hash it.  
         $category = lc $category;  
         # Use the category and tracelevel to compute the result.  
         if (ref $traceLevel) {  
             Confess("Bad trace level.");  
         } elsif (ref $TraceLevel) {  
             Confess("Bad trace config.");  
2570          }          }
         $retVal = ($traceLevel <= $TraceLevel && ($AllTrace || exists $Categories{$category}));  
2571      }      }
     # Return the computed result.  
     return $retVal;  
2572  }  }
2573    
2574  =head3 ParseCommand  =head3 SetPermissions
   
 C<< my ($options, @arguments) = Tracer::ParseCommand(\%optionTable, @inputList); >>  
2575    
2576  Parse a command line consisting of a list of parameters. The initial parameters may be option      Tracer::SetPermissions($dirName, $group, $mask, %otherMasks);
 specifiers of the form C<->I<option> or C<->I<option>C<=>I<value>. The options are stripped  
 off and merged into a table of default options. The remainder of the command line is  
 returned as a list of positional arguments. For example, consider the following invocation.  
2577    
2578  C<< my ($options, @arguments) = ParseCommand({ errors => 0, logFile => 'trace.log'}, @words); >>  Set the permissions for a directory and all the files and folders inside it.
2579    In addition, the group ownership will be changed to the specified value.
2580    
2581  In this case, the list @words will be treated as a command line. There are two options available,  This method is more vulnerable than most to permission and compatability
2582  B<errors> and B<logFile>. If @words has the following format  problems, so it does internal error recovery.
2583    
2584  C<< -logFile=error.log apple orange rutabaga >>  =over 4
2585    
2586  then at the end of the invocation, C<$options> will be  =item dirName
2587    
2588  C<< { errors => 0, logFile => 'error.log' } >>  Name of the directory to process.
2589    
2590  and C<@arguments> will contain  =item group
2591    
2592  C<< apple orange rutabaga >>  Name of the group to be assigned.
2593    
2594  The parser allows for some escape sequences. See L</UnEscape> for a description. There is no  =item mask
 support for quote characters.  
2595    
2596  =over 4  Permission mask. Bits that are C<1> in this mask will be ORed into the
2597    permission bits of any file or directory that does not already have them
2598    set to 1.
2599    
2600  =item optionTable  =item otherMasks
2601    
2602  Table of default options.  Map of search patterns to permission masks. If a directory name matches
2603    one of the patterns, that directory and all its members and subdirectories
2604    will be assigned the new pattern. For example, the following would
2605    assign 01664 to most files, but would use 01777 for directories named C<tmp>.
2606    
2607  =item inputList      Tracer::SetPermissions($dirName, 'fig', 01664, '^tmp$' => 01777);
2608    
2609  List of words on the command line.  The list is ordered, so the following would use 0777 for C<tmp1> and
2610    0666 for C<tmp>, C<tmp2>, or C<tmp3>.
2611    
2612  =item RETURN      Tracer::SetPermissions($dirName, 'fig', 01664, '^tmp1' => 0777,
2613                                                       '^tmp' => 0666);
2614    
2615  Returns a reference to the option table and a list of the positional arguments.  Note that the pattern matches are all case-insensitive, and only directory
2616    names are matched, not file names.
2617    
2618  =back  =back
2619    
2620  =cut  =cut
2621    
2622  sub ParseCommand {  sub SetPermissions {
2623      # Get the parameters.      # Get the parameters.
2624      my ($optionTable, @inputList) = @_;      my ($dirName, $group, $mask, @otherMasks) = @_;
2625      # Process any options in the input list.      # Set up for error recovery.
2626      my %overrides = ();      eval {
2627      while ((@inputList > 0) && ($inputList[0] =~ /^-/)) {          # Switch to the specified directory.
2628          # Get the current option.          ChDir($dirName);
2629          my $arg = shift @inputList;          # Get the group ID.
2630          # Pull out the option name.          my $gid = getgrnam($group);
2631          $arg =~ /^-([^=]*)/g;          # Get the mask for tracing.
2632          my $name = $1;          my $traceMask = sprintf("%04o", $mask) . "($mask)";
2633          # Check for an option value.          Trace("Fixing permissions for directory $dirName using group $group($gid) and mask $traceMask.") if T(File => 2);
2634          if ($arg =~ /\G=(.*)$/g) {          my $fixCount = 0;
2635              # Here we have a value for the option.          my $lookCount = 0;
2636              $overrides{$name} = UnEscape($1);          # @dirs will be a stack of directories to be processed.
2637            my @dirs = (getcwd());
2638            while (scalar(@dirs) > 0) {
2639                # Get the current directory.
2640                my $dir = pop @dirs;
2641                # Check for a match to one of the specified directory names. To do
2642                # that, we need to pull the individual part of the name off of the
2643                # whole path.
2644                my $simpleName = $dir;
2645                if ($dir =~ m!/([^/]+)$!) {
2646                    $simpleName = $1;
2647                }
2648                Trace("Simple directory name for $dir is $simpleName.") if T(File => 4);
2649                # Search for a match.
2650                my $match = 0;
2651                my $i;
2652                for ($i = 0; $i < $#otherMasks && ! $match; $i += 2) {
2653                    my $pattern = $otherMasks[$i];
2654                    if ($simpleName =~ /$pattern/i) {
2655                        $match = 1;
2656                    }
2657                }
2658                # Check for a match. Note we use $i-1 because the loop added 2
2659                # before terminating due to the match.
2660                if ($match && $otherMasks[$i-1] != $mask) {
2661                    # This directory matches one of the incoming patterns, and it's
2662                    # a different mask, so we process it recursively with that mask.
2663                    SetPermissions($dir, $group, $otherMasks[$i-1], @otherMasks);
2664          } else {          } else {
2665              # Here there is no value, so we use 1.                  # Here we can process normally. Get all of the non-hidden members.
2666              $overrides{$name} = 1;                  my @submems = OpenDir($dir, 1);
2667                    for my $submem (@submems) {
2668                        # Get the full name.
2669                        my $thisMem = "$dir/$submem";
2670                        Trace("Checking member $thisMem.") if T(4);
2671                        $lookCount++;
2672                        if ($lookCount % 1000 == 0) {
2673                            Trace("$lookCount members examined. Current is $thisMem. Mask is $traceMask") if T(File => 3);
2674          }          }
2675                        # Fix the group.
2676                        chown -1, $gid, $thisMem;
2677                        # Insure this member is not a symlink.
2678                        if (! -l $thisMem) {
2679                            # Get its info.
2680                            my $fileInfo = stat $thisMem;
2681                            # Only proceed if we got the info. Otherwise, it's a hard link
2682                            # and we want to skip it anyway.
2683                            if ($fileInfo) {
2684                                my $fileMode = $fileInfo->mode;
2685                                if (($fileMode & $mask) != $mask) {
2686                                    # Fix this member.
2687                                    $fileMode |= $mask;
2688                                    chmod $fileMode, $thisMem;
2689                                    $fixCount++;
2690      }      }
2691      # Merge the options into the defaults.                              # If it's a subdirectory, stack it.
2692      GetOptions($optionTable, \%overrides);                              if (-d $thisMem) {
2693      # Translate the remaining parameters.                                  push @dirs, $thisMem;
2694      my @retVal = ();                              }
2695      for my $inputParm (@inputList) {                          }
2696          push @retVal, UnEscape($inputParm);                      }
2697                    }
2698                }
2699            }
2700            Trace("$lookCount files and directories processed, $fixCount fixed.") if T(File => 2);
2701        };
2702        # Check for an error.
2703        if ($@) {
2704            Confess("SetPermissions error: $@");
2705      }      }
     # Return the results.  
     return ($optionTable, @retVal);  
2706  }  }
2707    
2708  =head3 Escape  =head3 GetLine
2709    
2710  C<< my $codedString = Tracer::Escape($realString); >>      my @data = Tracer::GetLine($handle);
2711    
2712  Escape a string for use in a command length. Tabs will be replaced by C<\t>, new-lines  Read a line of data from a tab-delimited file.
 replaced by C<\n>, carriage returns will be deleted, and backslashes will be doubled. The  
 result is to reverse the effect of L</UnEscape>.  
2713    
2714  =over 4  =over 4
2715    
2716  =item realString  =item handle
2717    
2718  String to escape.  Open file handle from which to read.
2719    
2720  =item RETURN  =item RETURN
2721    
2722  Escaped equivalent of the real string.  Returns a list of the fields in the record read. The fields are presumed to be
2723    tab-delimited. If we are at the end of the file, then an empty list will be
2724    returned. If an empty line is read, a single list item consisting of a null
2725    string will be returned.
2726    
2727  =back  =back
2728    
2729  =cut  =cut
2730    
2731  sub Escape {  sub GetLine {
2732      # Get the parameter.      # Get the parameters.
2733      my ($realString) = @_;      my ($handle) = @_;
2734      # Initialize the return variable.      # Declare the return variable.
2735      my $retVal = "";      my @retVal = ();
2736      # Loop through the parameter string, looking for sequences to escape.      Trace("File position is " . tell($handle) . ". EOF flag is " . eof($handle) . ".") if T(File => 4);
2737      while (length $realString > 0) {      # Read from the file.
2738          # Look for the first sequence to escape.      my $line = <$handle>;
2739          if ($realString =~ /^(.*?)([\n\t\r\\])/) {      # Only proceed if we found something.
2740              # Here we found it. The text preceding the sequence is in $1. The sequence      if (defined $line) {
2741              # itself is in $2. First, move the clear text to the return variable.          # Remove the new-line. We are a bit over-cautious here because the file may be coming in via an
2742              $retVal .= $1;          # upload control and have a nonstandard EOL combination.
2743              # Strip the processed section off the real string.          $line =~ s/(\r|\n)+$//;
2744              $realString = substr $realString, (length $2) + (length $1);          # Here we do some fancy tracing to help in debugging complicated EOL marks.
2745              # Get the matched character.          if (T(File => 4)) {
2746              my $char = $2;              my $escapedLine = $line;
2747              # If we have a CR, we are done.              $escapedLine =~ s/\n/\\n/g;
2748              if ($char ne "\r") {              $escapedLine =~ s/\r/\\r/g;
2749                  # It's not a CR, so encode the escape sequence.              $escapedLine =~ s/\t/\\t/g;
2750                  $char =~ tr/\t\n/tn/;              Trace("Line read: -->$escapedLine<--");
2751                  $retVal .= "\\" . $char;          }
2752              }          # If the line is empty, return a single empty string; otherwise, parse
2753            # it into fields.
2754            if ($line eq "") {
2755                push @retVal, "";
2756          } else {          } else {
2757              # Here there are no more escape sequences. The rest of the string is              push @retVal, split /\t/,$line;
             # transferred unmodified.  
             $retVal .= $realString;  
             $realString = "";  
2758          }          }
2759        } else {
2760            # Trace the reason the read failed.
2761            Trace("End of file: $!") if T(File => 3);
2762      }      }
2763      # Return the result.      # Return the result.
2764      return $retVal;      return @retVal;
2765  }  }
2766    
2767  =head3 UnEscape  =head3 PutLine
2768    
2769  C<< my $realString = Tracer::UnEscape($codedString); >>      Tracer::PutLine($handle, \@fields, $eol);
2770    
2771  Replace escape sequences with their actual equivalents. C<\t> will be replaced by  Write a line of data to a tab-delimited file. The specified field values will be
2772  a tab, C<\n> by a new-line character, and C<\\> by a backslash. C<\r> codes will  output in tab-separated form, with a trailing new-line.
 be deleted.  
2773    
2774  =over 4  =over 4
2775    
2776  =item codedString  =item handle
2777    
2778  String to un-escape.  Output file handle.
2779    
2780  =item RETURN  =item fields
2781    
2782  Returns a copy of the original string with the escape sequences converted to their actual  List of field values.
2783  values.  
2784    =item eol (optional)
2785    
2786    End-of-line character (default is "\n").
2787    
2788  =back  =back
2789    
2790  =cut  =cut
2791    
2792  sub UnEscape {  sub PutLine {
2793      # Get the parameter.      # Get the parameters.
2794      my ($codedString) = @_;      my ($handle, $fields, $eol) = @_;
2795      # Initialize the return variable.      # Write the data.
2796      my $retVal = "";      print $handle join("\t", @{$fields}) . ($eol || "\n");
     # Only proceed if the incoming string is nonempty.  
     if (defined $codedString) {  
         # Loop through the parameter string, looking for escape sequences. We can't do  
         # translating because it causes problems with the escaped slash. ("\\t" becomes  
         # "\<tab>" no matter what we do.)  
         while (length $codedString > 0) {  
             # Look for the first escape sequence.  
             if ($codedString =~ /^(.*?)\\(\\|n|t|r)/) {  
                 # Here we found it. The text preceding the sequence is in $1. The sequence  
                 # itself is in $2. First, move the clear text to the return variable.  
                 $retVal .= $1;  
                 $codedString = substr $codedString, (2 + length $1);  
                 # Get the escape value.  
                 my $char = $2;  
                 # If we have a "\r", we are done.  
                 if ($char ne 'r') {  
                     # Here it's not an 'r', so we convert it.  
                     $char =~ tr/\\tn/\\\t\n/;  
                     $retVal .= $char;  
                 }  
             } else {  
                 # Here there are no more escape sequences. The rest of the string is  
                 # transferred unmodified.  
                 $retVal .= $codedString;  
                 $codedString = "";  
             }  
         }  
     }  
     # Return the result.  
     return $retVal;  
2797  }  }
2798    
 =head3 ParseRecord  
2799    
2800  C<< my @fields = Tracer::ParseRecord($line); >>  =head3 PrintLine
2801    
2802  Parse a tab-delimited data line. The data line is split into field values. Embedded tab      Tracer::PrintLine($line);
2803  and new-line characters in the data line must be represented as C<\t> and C<\n>, respectively.  
2804  These will automatically be converted.  Print a line of text with a trailing new-line.
2805    
2806  =over 4  =over 4
2807    
2808  =item line  =item line
2809    
2810  Line of data containing the tab-delimited fields.  Line of text to print.
   
 =item RETURN  
   
 Returns a list of the fields found in the data line.  
2811    
2812  =back  =back
2813    
2814  =cut  =cut
2815    
2816  sub ParseRecord {  sub PrintLine {
2817      # Get the parameter.      # Get the parameters.
2818      my ($line) = @_;      my ($line) = @_;
2819      # Remove the trailing new-line, if any.      # Print the line.
2820      chomp $line;      print "$line\n";
     # Split the line read into pieces using the tab character.  
     my @retVal = split /\t/, $line;  
     # Trim and fix the escapes in each piece.  
     for my $value (@retVal) {  
         # Trim leading whitespace.  
         $value =~ s/^\s+//;  
         # Trim trailing whitespace.  
         $value =~ s/\s+$//;  
         # Delete the carriage returns.  
         $value =~ s/\r//g;  
         # Convert the escapes into their real values.  
         $value =~ s/\\t/"\t"/ge;  
         $value =~ s/\\n/"\n"/ge;  
     }  
     # Return the result.  
     return @retVal;  
2821  }  }
2822    
 =head3 Merge  
2823    
2824  C<< my @mergedList = Tracer::Merge(@inputList); >>  =head2 Other Useful Methods
2825    
2826  Sort a list of strings and remove duplicates.  =head3 ParseParm
2827    
2828        my $listValue = Tracer::ParseParm($string);
2829    
2830    Convert a parameter into a list reference. If the parameter is undefined,
2831    an undefined value will be returned. Otherwise, it will be parsed as a
2832    comma-separated list of values.
2833    
2834  =over 4  =over 4
2835    
2836  =item inputList  =item string
2837    
2838  List of scalars to sort and merge.  Incoming string.
2839    
2840  =item RETURN  =item RETURN
2841    
2842  Returns a list containing the same elements sorted in ascending order with duplicates  Returns a reference to a list of values, or C<undef> if the incoming value
2843  removed.  was undefined.
2844    
2845  =back  =back
2846    
2847  =cut  =cut
2848    
2849  sub Merge {  sub ParseParm {
2850      # Get the input list in sort order.      # Get the parameters.
2851      my @inputList = sort @_;      my ($string) = @_;
2852      # Only proceed if the list has at least two elements.      # Declare the return variable.
2853      if (@inputList > 1) {      my $retVal;
2854          # Now we want to move through the list splicing out duplicates.      # Check for data.
2855          my $i = 0;      if (defined $string) {
2856          while ($i < @inputList) {          # We have some, so split it into a list.
2857              # Get the current entry.          $retVal = [ split /\s*,\s*/, $string];
             my $thisEntry = $inputList[$i];  
             # Find out how many elements duplicate the current entry.  
             my $j = $i + 1;  
             my $dup1 = $i + 1;  
             while ($j < @inputList && $inputList[$j] eq $thisEntry) { $j++; };  
             # If the number is nonzero, splice out the duplicates found.  
             if ($j > $dup1) {  
                 splice @inputList, $dup1, $j - $dup1;  
             }  
             # Now the element at position $dup1 is different from the element before it  
             # at position $i. We push $i forward one position and start again.  
             $i++;  
         }  
2858      }      }
2859      # Return the merged list.      # Return the result.
2860      return @inputList;      return $retVal;
2861  }  }
2862    
2863  =head3 GetFile  =head3 Now
2864    
2865  C<< my @fileContents = Tracer::GetFile($fileName); >>      my $string = Tracer::Now();
2866    
2867      or  Return a displayable time stamp containing the local time. Whatever format this
2868    method produces must be parseable by L</ParseDate>.
2869    
2870  C<< my $fileContents = Tracer::GetFile($fileName); >>  =cut
2871    
2872  Return the entire contents of a file. In list context, line-ends are removed and  sub Now {
2873  each line is a list element. In scalar context, line-ends are replaced by C<\n>.      return DisplayTime(time);
2874    }
2875    
2876    =head3 DisplayTime
2877    
2878        my $string = Tracer::DisplayTime($time);
2879    
2880    Convert a time value to a displayable time stamp. Whatever format this
2881    method produces must be parseable by L</ParseDate>.
2882    
2883  =over 4  =over 4
2884    
2885  =item fileName  =item time
2886    
2887  Name of the file to read.  Time to display, in seconds since the epoch, or C<undef> if the time is unknown.
2888    
2889  =item RETURN  =item RETURN
2890    
2891  In a list context, returns the entire file as a list with the line terminators removed.  Returns a displayable time, or C<(n/a)> if the incoming time is undefined.
 In a scalar context, returns the entire file as a string. If an error occurs opening  
 the file, an empty list will be returned.  
2892    
2893  =back  =back
2894    
2895  =cut  =cut
2896    
2897  sub GetFile {  sub DisplayTime {
2898      # Get the parameters.      my ($time) = @_;
2899      my ($fileName) = @_;      my $retVal = "(n/a)";
2900      # Declare the return variable.      if (defined $time) {
2901      my @retVal = ();          my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime($time);
2902      # Open the file for input.          $retVal = _p2($mon+1) . "/" . _p2($mday) . "/" . ($year + 1900) . " " .
2903      my $ok = open INPUTFILE, "<$fileName";                    _p2($hour) . ":" . _p2($min) . ":" . _p2($sec);
     if (!$ok) {  
         # If we had an error, trace it. We will automatically return a null value.  
         Trace("Could not open \"$fileName\" for input: $!") if T(0);  
     } else {  
         # Read the whole file into the return variable, stripping off any terminator  
         # characters.  
         my $lineCount = 0;  
         while (my $line = <INPUTFILE>) {  
             $lineCount++;  
             $line = Strip($line);  
             push @retVal, $line;  
         }  
         # Close it.  
         close INPUTFILE;  
         my $actualLines = @retVal;  
2904      }      }
2905      # Return the file's contents in the desired format.      return $retVal;
     if (wantarray) {  
         return @retVal;  
     } else {  
         return join "\n", @retVal;  
2906      }      }
2907    
2908    # Pad a number to 2 digits.
2909    sub _p2 {
2910        my ($value) = @_;
2911        $value = "0$value" if ($value < 10);
2912        return $value;
2913  }  }
2914    
2915  =head3 QTrace  =head3 Escape
2916    
2917  C<< my $data = QTrace($format); >>      my $codedString = Tracer::Escape($realString);
2918    
2919  Return the queued trace data in the specified format.  Escape a string for use in a command. Tabs will be replaced by C<\t>, new-lines
2920    replaced by C<\n>, carriage returns will be deleted, and backslashes will be doubled. The
2921    result is to reverse the effect of L</UnEscape>.
2922    
2923  =over 4  =over 4
2924    
2925  =item format  =item realString
2926    
2927  C<html> to format the data as an HTML list, C<text> to format it as straight text.  String to escape.
2928    
2929    =item RETURN
2930    
2931    Escaped equivalent of the real string.
2932    
2933  =back  =back
2934    
2935  =cut  =cut
2936    
2937  sub QTrace {  sub Escape {
2938      # Get the parameter.      # Get the parameter.
2939      my ($format) = @_;      my ($realString) = @_;
2940      # Create the return variable.      # Initialize the return variable.
2941      my $retVal = "";      my $retVal = "";
2942      # Only proceed if there is an actual queue.      # Loop through the parameter string, looking for sequences to escape.
2943      if (@Queue) {      while (length $realString > 0) {
2944          # Process according to the format.          # Look for the first sequence to escape.
2945          if ($format =~ m/^HTML$/i) {          if ($realString =~ /^(.*?)([\n\t\r\\])/) {
2946              # Convert the queue into an HTML list.              # Here we found it. The text preceding the sequence is in $1. The sequence
2947              $retVal = "<ul>\n";              # itself is in $2. First, move the clear text to the return variable.
2948              for my $line (@Queue) {              $retVal .= $1;
2949                  my $escapedLine = CGI::escapeHTML($line);              # Strip the processed section off the real string.
2950                  $retVal .= "<li>$escapedLine</li>\n";              $realString = substr $realString, (length $2) + (length $1);
2951                # Get the matched character.
2952                my $char = $2;
2953                # If we have a CR, we are done.
2954                if ($char ne "\r") {
2955                    # It's not a CR, so encode the escape sequence.
2956                    $char =~ tr/\t\n/tn/;
2957                    $retVal .= "\\" . $char;
2958              }              }
2959              $retVal .= "</ul>\n";          } else {
2960          } elsif ($format =~ m/^TEXT$/i) {              # Here there are no more escape sequences. The rest of the string is
2961              # Convert the queue into a list of text lines.              # transferred unmodified.
2962              $retVal = join("\n", @Queue) . "\n";              $retVal .= $realString;
2963                $realString = "";
2964          }          }
         # Clear the queue.  
         @Queue = ();  
2965      }      }
2966      # Return the formatted list.      # Return the result.
2967      return $retVal;      return $retVal;
2968  }  }
2969    
2970  =head3 Confess  =head3 UnEscape
2971    
2972  C<< Confess($message); >>      my $realString = Tracer::UnEscape($codedString);
2973    
2974  Trace the call stack and abort the program with the specified message. When used with  Replace escape sequences with their actual equivalents. C<\t> will be replaced by
2975  the OR operator and the L</Assert> method, B<Confess> can function as a debugging assert.  a tab, C<\n> by a new-line character, and C<\\> by a backslash. C<\r> codes will
2976  So, for example  be deleted.
2977    
2978  C<< Assert($recNum >= 0) || Confess("Invalid record number $recNum."); >>  =over 4
2979    
2980  Will abort the program with a stack trace if the value of C<$recNum> is negative.  =item codedString
2981    
2982  =over 4  String to un-escape.
2983    
2984  =item message  =item RETURN
2985    
2986  Message to include in the trace.  Returns a copy of the original string with the escape sequences converted to their actual
2987    values.
2988    
2989  =back  =back
2990    
2991  =cut  =cut
2992    
2993  sub Confess {  sub UnEscape {
2994      # Get the parameters.      # Get the parameter.
2995      my ($message) = @_;      my ($codedString) = @_;
2996      # Trace the call stack.      # Initialize the return variable.
2997      Cluck($message);      my $retVal = "";
2998      # Abort the program.      # Only proceed if the incoming string is nonempty.
2999      croak(">>> $message");      if (defined $codedString) {
3000            # Loop through the parameter string, looking for escape sequences. We can't do
3001            # translating because it causes problems with the escaped slash. ("\\t" becomes
3002            # "\<tab>" no matter what we do.)
3003            while (length $codedString > 0) {
3004                # Look for the first escape sequence.
3005                if ($codedString =~ /^(.*?)\\(\\|n|t|r)/) {
3006                    # Here we found it. The text preceding the sequence is in $1. The sequence
3007                    # itself is in $2. First, move the clear text to the return variable.
3008                    $retVal .= $1;
3009                    $codedString = substr $codedString, (2 + length $1);
3010                    # Get the escape value.
3011                    my $char = $2;
3012                    # If we have a "\r", we are done.
3013                    if ($char ne 'r') {
3014                        # Here it's not an 'r', so we convert it.
3015                        $char =~ tr/\\tn/\\\t\n/;
3016                        $retVal .= $char;
3017                    }
3018                } else {
3019                    # Here there are no more escape sequences. The rest of the string is
3020                    # transferred unmodified.
3021                    $retVal .= $codedString;
3022                    $codedString = "";
3023                }
3024            }
3025        }
3026        # Return the result.
3027        return $retVal;
3028  }  }
3029    
3030  =head3 Assert  =head3 Percent
3031    
3032  C<< Assert($condition1, $condition2, ... $conditionN); >>      my $percent = Tracer::Percent($number, $base);
3033    
3034  Return TRUE if all the conditions are true. This method can be used in conjunction with  Returns the percent of the base represented by the given number. If the base
3035  the OR operator and the L</Confess> method as a debugging assert.  is zero, returns zero.
3036  So, for example  
3037    =over 4
3038    
3039  C<< Assert($recNum >= 0) || Confess("Invalid record number $recNum."); >>  =item number
3040    
3041  Will abort the program with a stack trace if the value of C<$recNum> is negative.  Percent numerator.
3042    
3043    =item base
3044    
3045    Percent base.
3046    
3047    =item RETURN
3048    
3049    Returns the percentage of the base represented by the numerator.
3050    
3051    =back
3052    
3053  =cut  =cut
3054  sub Assert {  
3055      my $retVal = 1;  sub Percent {
3056      LOOP: for my $condition (@_) {      # Get the parameters.
3057          if (! $condition) {      my ($number, $base) = @_;
3058              $retVal = 0;      # Declare the return variable.
3059              last LOOP;      my $retVal = 0;
3060          }      # Compute the percent.
3061        if ($base != 0) {
3062            $retVal = $number * 100 / $base;
3063      }      }
3064        # Return the result.
3065      return $retVal;      return $retVal;
3066  }  }
3067    
3068  =head3 Cluck  =head3 Constrain
3069    
3070  C<< Cluck($message); >>      my $constrained = Constrain($value, $min, $max);
3071    
3072  Trace the call stack. Note that for best results, you should qualify the call with a  Modify a numeric value to bring it to a point in between a maximum and a minimum.
 trace condition. For example,  
3073    
3074  C<< Cluck("Starting record parse.") if T(3); >>  =over 4
3075    
3076  will only trace the stack if the trace level for the package is 3 or more.  =item value
3077    
3078  =over 4  Value to constrain.
3079    
3080  =item message  =item min (optional)
3081    
3082  Message to include in the trace.  Minimum permissible value. If this parameter is undefined, no minimum constraint will be applied.
3083    
3084    =item max (optional)
3085    
3086    Maximum permissible value. If this parameter is undefined, no maximum constraint will be applied.
3087    
3088    =item RETURN
3089    
3090    Returns the incoming value, constrained according to the other parameters.
3091    
3092  =back  =back
3093    
3094  =cut  =cut
3095    
3096  sub Cluck {  sub Constrain {
3097      # Get the parameters.      # Get the parameters.
3098      my ($message) = @_;      my ($value, $min, $max) = @_;
3099      # Trace what's happening.      # Declare the return variable.
3100      Trace("Stack trace for event: $message");      my $retVal = $value;
3101      my $confession = longmess($message);      # Apply the minimum constraint.
3102      # Convert the confession to a series of trace messages. Note we skip any      if (defined $min && $retVal < $min) {
3103      # messages relating to calls into Tracer.          $retVal = $min;
3104      for my $line (split /\s*\n/, $confession) {      }
3105          Trace($line) if ($line !~ /Tracer\.pm/);      # Apply the maximum constraint.
3106        if (defined $max && $retVal > $max) {
3107            $retVal = $max;
3108      }      }
3109        # Return the result.
3110        return $retVal;
3111  }  }
3112    
3113  =head3 Min  =head3 Min
3114    
3115  C<< my $min = Min($value1, $value2, ... $valueN); >>      my $min = Min($value1, $value2, ... $valueN);
3116    
3117  Return the minimum argument. The arguments are treated as numbers.  Return the minimum argument. The arguments are treated as numbers.
3118    
# Line 1573  Line 3145 
3145    
3146  =head3 Max  =head3 Max
3147    
3148  C<< my $max = Max($value1, $value2, ... $valueN); >>      my $max = Max($value1, $value2, ... $valueN);
3149    
3150  Return the maximum argument. The arguments are treated as numbers.  Return the maximum argument. The arguments are treated as numbers.
3151    
# Line 1604  Line 3176 
3176      return $retVal;      return $retVal;
3177  }  }
3178    
 =head3 AddToListMap  
   
 C<< Tracer::AddToListMap(\%hash, $key, $value); >>  
   
 Add a key-value pair to a hash of lists. If no value exists for the key, a singleton list  
 is created for the key. Otherwise, the new value is pushed onto the list.  
   
 =over 4  
   
 =item hash  
   
 Reference to the target hash.  
   
 =item key  
   
 Key for which the value is to be added.  
   
 =item value  
   
 Value to add to the key's value list.  
   
 =back  
   
 =cut  
   
 sub AddToListMap {  
     # Get the parameters.  
     my ($hash, $key, $value) = @_;  
     # Process according to whether or not the key already has a value.  
     if (! exists $hash->{$key}) {  
         $hash->{$key} = [$value];  
     } else {  
         push @{$hash->{$key}}, $value;  
     }  
 }  
   
 =head3 DebugMode  
   
 C<< if (Tracer::DebugMode) { ...code... } >>  
   
 Return TRUE if debug mode has been turned on, else output an error  
 page and return FALSE.  
   
 Certain CGI scripts are too dangerous to exist in the production  
 environment. This method provides a simple way to prevent them  
 from working unless they are explicitly turned on by creating a password  
 cookie via the B<SetPassword> script.  If debugging mode  
 is not turned on, an error web page will be output directing the  
 user to enter in the correct password.  
   
 =cut  
   
 sub DebugMode {  
     # Declare the return variable.  
     my $retVal = 0;  
     # Check the debug configuration.  
     my $password = CGI::cookie("DebugMode");  
     my $encrypted = Digest::MD5::md5_hex($password);  
     if ($encrypted eq "252dec43280e0c0d6a75ffcec486e61d") {  
         $retVal = 1;  
     } else {  
         # Here debug mode is off, so we generate an error page.  
         my $pageString = PageBuilder::Build("<Html/ErrorPage.html", {}, "Html");  
         print $pageString;  
     }  
     # Return the determination indicator.  
     return $retVal;  
 }  
3179    
3180  =head3 Strip  =head3 Strip
3181    
3182  C<< my $string = Tracer::Strip($line); >>      my $string = Tracer::Strip($line);
3183    
3184  Strip all line terminators off a string. This is necessary when dealing with files  Strip all line terminators off a string. This is necessary when dealing with files
3185  that may have been transferred back and forth several times among different  that may have been transferred back and forth several times among different
# Line 1707  Line 3211 
3211    
3212  =head3 Pad  =head3 Pad
3213    
3214  C<< my $paddedString = Tracer::Pad($string, $len, $left, $padChar); >>      my $paddedString = Tracer::Pad($string, $len, $left, $padChar);
3215    
3216  Pad a string to a specified length. The pad character will be a  Pad a string to a specified length. The pad character will be a
3217  space, and the padding will be on the right side unless specified  space, and the padding will be on the right side unless specified
# Line 1778  Line 3282 
3282    
3283  =head3 TICK  =head3 TICK
3284    
3285  C<< my @results = TICK($commandString); >>      my @results = TICK($commandString);
3286    
3287  Perform a back-tick operation on a command. If this is a Windows environment, any leading  Perform a back-tick operation on a command. If this is a Windows environment, any leading
3288  dot-slash (C<./> will be removed. So, for example, if you were doing  dot-slash (C<./> will be removed. So, for example, if you were doing
# Line 1817  Line 3321 
3321      return `$commandString`;      return `$commandString`;
3322  }  }
3323    
 =head3 ScriptSetup  
3324    
3325  C<< my ($query, $varHash) = ScriptSetup(); >>  =head3 CommaFormat
3326    
3327        my $formatted = Tracer::CommaFormat($number);
3328    
3329    Insert commas into a number.
3330    
3331    =over 4
3332    
3333    =item number
3334    
3335    A sequence of digits.
3336    
3337  Perform standard tracing and debugging setup for scripts. The value returned is  =item RETURN
3338  the CGI object followed by a pre-built variable hash.  
3339    Returns the same digits with commas strategically inserted.
3340    
3341  The C<Trace> query parameter is used to determine whether or not tracing is active and  =back
 which trace modules (other than C<Tracer> and C<FIG>) should be turned on. Specifying  
 the C<CGI> trace module will trace parameter and environment information. Parameters are  
 traced at level 3 and environment variables at level 4. At the end of the script, the  
 client should call L</ScriptFinish> to output the web page.  
3342    
3343  =cut  =cut
3344    
3345  sub ScriptSetup {  sub CommaFormat {
3346      # Get the CGI query object.      # Get the parameters.
3347      my $query = CGI->new();      my ($number) = @_;
3348      # Check for tracing. Set it up if the user asked for it.      # Pad the length up to a multiple of three.
3349      if ($query->param('Trace')) {      my $padded = "$number";
3350          # Set up tracing to be queued for display at the bottom of the web page.      $padded = " " . $padded while length($padded) % 3 != 0;
3351          TSetup($query->param('Trace') . " FIG Tracer", "QUEUE");      # This is a fancy PERL trick. The parentheses in the SPLIT pattern
3352          # Trace the parameter and environment data.      # cause the delimiters to be included in the output stream. The
3353          if (T(CGI => 3)) {      # GREP removes the empty strings in between the delimiters.
3354              # Here we want to trace the parameter data.      my $retVal = join(",", grep { $_ ne '' } split(/(...)/, $padded));
3355              my @names = $query->param;      # Clean out the spaces.
3356              for my $parmName (sort @names) {      $retVal =~ s/ //g;
3357                  # Note we skip "Trace", which is for our use only.      # Return the result.
3358                  if ($parmName ne 'Trace') {      return $retVal;
                     my @values = $query->param($parmName);  
                     Trace("CGI: $parmName = " . join(", ", @values));  
                 }  
             }  
         }  
         if (T(CGI => 4)) {  
             # Here we want the environment data too.  
             for my $envName (sort keys %ENV) {  
                 Trace("ENV: $envName = $ENV{$envName}");  
             }  
         }  
     } else {  
         # Here tracing is to be turned off. All we allow is errors traced into the  
         # error log.  
         TSetup("0", "WARN");  
     }  
     # Create the variable hash.  
     my $varHash = { DebugData => '' };  
     # If we're in DEBUG mode, set up the debug mode data for forms.  
     if (Tracer::DebugMode) {  
         $varHash->{DebugData} = GetFile("Html/DebugFragment.html");  
     }  
     # Return the query object and variable hash.  
     return ($query, $varHash);  
3359  }  }
3360    
 =head3 ScriptFinish  
3361    
3362  C<< ScriptFinish($webData, $varHash); >>  =head3 CompareLists
3363    
3364  Output a web page at the end of a script. Either the string to be output or the      my ($inserted, $deleted) = Tracer::CompareLists(\@newList, \@oldList, $keyIndex);
 name of a template file can be specified. If the second parameter is omitted,  
 it is assumed we have a string to be output; otherwise, it is assumed we have the  
 name of a template file. The template should have the variable C<DebugData>  
 specified in any form that invokes a standard script. If debugging mode is turned  
 on, a form field will be put in that allows the user to enter tracing data.  
 Trace messages will be placed immediately before the terminal C<BODY> tag in  
 the output, formatted as a list.  
3365    
3366  A typical standard script would loook like the following.  Compare two lists of tuples, and return a hash analyzing the differences. The lists
3367    are presumed to be sorted alphabetically by the value in the $keyIndex column.
3368    The return value contains a list of items that are only in the new list
3369    (inserted) and only in the old list (deleted).
3370    
3371      BEGIN {  =over 4
         # Print the HTML header.  
         print "CONTENT-TYPE: text/html\n\n";  
     }  
     use Tracer;  
     use CGI;  
     use FIG;  
     # ... more uses ...  
3372    
3373      my ($query, $varHash) = ScriptSetup();  =item newList
     eval {  
         # ... get data from $query, put it in $varHash ...  
     };  
     if ($@) {  
         Trace("Script Error: $@") if T(0);  
     }  
     ScriptFinish("Html/MyTemplate.html", $varHash);  
3374    
3375  The idea here is that even if the script fails, you'll see trace messages and  Reference to a list of new tuples.
 useful output.  
3376    
3377  =over 4  =item oldList
3378    
3379  =item webData  Reference to a list of old tuples.
3380    
3381  A string containing either the full web page to be written to the output or the  =item keyIndex (optional)
 name of a template file from which the page is to be constructed. If the name  
 of a template file is specified, then the second parameter must be present;  
 otherwise, it must be absent.  
3382    
3383  =item varHash (optional)  Index into each tuple of its key field. The default is 0.
3384    
3385  If specified, then a reference to a hash mapping variable names for a template  =item RETURN
3386  to their values. The template file will be read into memory, and variable markers  
3387  will be replaced by data in this hash reference.  Returns a 2-tuple consisting of a reference to the list of items that are only in the new
3388    list (inserted) followed by a reference to the list of items that are only in the old
3389    list (deleted).
3390    
3391  =back  =back
3392    
3393  =cut  =cut
3394    
3395  sub ScriptFinish {  sub CompareLists {
3396      # Get the parameters.      # Get the parameters.
3397      my ($webData, $varHash) = @_;      my ($newList, $oldList, $keyIndex) = @_;
3398      # Check for a template file situation.      if (! defined $keyIndex) {
3399      my $outputString;          $keyIndex = 0;
3400      if (defined $varHash) {      }
3401          # Here we have a template file. We need to apply the variables to the template.      # Declare the return variables.
3402          $outputString = PageBuilder::Build("<$webData", $varHash, "Html");      my ($inserted, $deleted) = ([], []);
3403      } else {      # Loop through the two lists simultaneously.
3404          # Here the user gave us a raw string.      my ($newI, $oldI) = (0, 0);
3405          $outputString = $webData;      my ($newN, $oldN) = (scalar @{$newList}, scalar @{$oldList});
3406      }      while ($newI < $newN || $oldI < $oldN) {
3407      # Check for trace messages.          # Get the current object in each list. Note that if one
3408      if ($Destination eq "QUEUE") {          # of the lists is past the end, we'll get undef.
3409          # We have trace messages, so we want to put them at the end of the body. This          my $newItem = $newList->[$newI];
3410          # is either at the end of the whole string or at the beginning of the BODY          my $oldItem = $oldList->[$oldI];
3411          # end-tag.          if (! defined($newItem) || defined($oldItem) && $newItem->[$keyIndex] gt $oldItem->[$keyIndex]) {
3412          my $pos = length $outputString;              # The old item is not in the new list, so mark it deleted.
3413          if ($outputString =~ m#</body>#gi) {              push @{$deleted}, $oldItem;
3414              $pos = (pos $outputString) - 7;              $oldI++;
3415            } elsif (! defined($oldItem) || $oldItem->[$keyIndex] gt $newItem->[$keyIndex]) {
3416                # The new item is not in the old list, so mark it inserted.
3417                push @{$inserted}, $newItem;
3418                $newI++;
3419            } else {
3420                # The item is in both lists, so push forward.
3421                $oldI++;
3422                $newI++;
3423          }          }
         substr $outputString, $pos, 0, QTrace('Html');  
3424      }      }
3425      # Write the output string.      # Return the result.
3426      print $outputString;      return ($inserted, $deleted);
3427  }  }
3428    
3429  =head3 Insure  =head3 GenerateURL
3430    
3431  C<< Insure($dirName); >>      my $queryUrl = Tracer::GenerateURL($page, %parameters);
3432    
3433  Insure a directory is present.  Generate a GET-style URL for the specified page with the specified parameter
3434    names and values. The values will be URL-escaped automatically. So, for
3435    example
3436    
3437        Tracer::GenerateURL("form.cgi", type => 1, string => "\"high pass\" or highway")
3438    
3439    would return
3440    
3441        form.cgi?type=1;string=%22high%20pass%22%20or%20highway
3442    
3443  =over 4  =over 4
3444    
3445  =item dirName  =item page
3446    
3447  Name of the directory to check. If it does not exist, it will be created.  Page URL.
3448    
3449    =item parameters
3450    
3451    Hash mapping parameter names to parameter values.
3452    
3453    =item RETURN
3454    
3455    Returns a GET-style URL that goes to the specified page and passes in the
3456    specified parameters and values.
3457    
3458  =back  =back
3459    
3460  =cut  =cut
3461    
3462  sub Insure {  sub GenerateURL {
3463      my ($dirName) = @_;      # Get the parameters.
3464      if (! -d $dirName) {      my ($page, %parameters) = @_;
3465          Trace("Creating $dirName directory.") if T(2);      # Prime the return variable with the page URL.
3466          eval { mkpath $dirName; };      my $retVal = $page;
3467          if ($@) {      # Loop through the parameters, creating parameter elements in a list.
3468              Confess("Error creating $dirName: $@");      my @parmList = map { "$_=" . uri_escape($parameters{$_}) } keys %parameters;
3469          }      # If the list is nonempty, tack it on.
3470        if (@parmList) {
3471            $retVal .= "?" . join(";", @parmList);
3472      }      }
3473        # Return the result.
3474        return $retVal;
3475  }  }
3476    
3477  =head3 ChDir  =head3 ApplyURL
3478    
3479  C<< ChDir($dirName); >>      Tracer::ApplyURL($table, $target, $url);
3480    
3481  Change to the specified directory.  Run through a two-dimensional table (or more accurately, a list of lists), converting the
3482    I<$target> column to HTML text having a hyperlink to a URL in the I<$url> column. The
3483    URL column will be deleted by this process and the target column will be HTML-escaped.
3484    
3485    This provides a simple way to process the results of a database query into something
3486    displayable by combining a URL with text.
3487    
3488  =over 4  =over 4
3489    
3490  =item dirName  =item table
3491    
3492  Name of the directory to which we want to change.  Reference to a list of lists. The elements in the containing list will be updated by
3493    this method.
3494    
3495    =item target
3496    
3497    The index of the column to be converted into HTML.
3498    
3499    =item url
3500    
3501    The index of the column containing the URL. Note that the URL must have a recognizable
3502    C<http:> at the beginning.
3503    
3504  =back  =back
3505    
3506  =cut  =cut
3507    
3508  sub ChDir {  sub ApplyURL {
3509      my ($dirName) = @_;      # Get the parameters.
3510      if (! -d $dirName) {      my ($table, $target, $url) = @_;
3511          Confess("Cannot change to directory $dirName: no such directory.");      # Loop through the table.
3512      } else {      for my $row (@{$table}) {
3513          Trace("Changing to directory $dirName.") if T(4);          # Apply the URL to the target cell.
3514          my $okFlag = chdir $dirName;          $row->[$target] = CombineURL($row->[$target], $row->[$url]);
3515          if (! $okFlag) {          # Delete the URL from the row.
3516              Confess("Error switching to directory $dirName.");          delete $row->[$url];
         }  
3517      }      }
3518  }  }
3519    
3520  =head3 SetPermissions  =head3 CombineURL
   
 C<< Tracer::SetPermissions($dirName, $group, $mask, %otherMasks); >>  
3521    
3522  Set the permissions for a directory and all the files and folders inside it.      my $combinedHtml = Tracer::CombineURL($text, $url);
 In addition, the group ownership will be changed to the specified value.  
3523    
3524  This method is more vulnerable than most to permission and compatability  This method will convert the specified text into HTML hyperlinked to the specified
3525  problems, so it does internal error recovery.  URL. The hyperlinking will only take place if the URL looks legitimate: that is, it
3526    is defined and begins with an C<http:> header.
3527    
3528  =over 4  =over 4
3529    
3530  =item dirName  =item text
3531    
3532  Name of the directory to process.  Text to return. This will be HTML-escaped automatically.
3533    
3534  =item group  =item url
3535    
3536  Name of the group to be assigned.  A URL to be hyperlinked to the text. If it does not look like a URL, then the text
3537    will be returned without any hyperlinking.
3538    
3539  =item mask  =item RETURN
3540    
3541  Permission mask. Bits that are C<1> in this mask will be ORed into the  Returns the original text, HTML-escaped, with the URL hyperlinked to it. If the URL
3542  permission bits of any file or directory that does not already have them  doesn't look right, the HTML-escaped text will be returned without any further
3543  set to 1.  modification.
3544    
3545  =item otherMasks  =back
3546    
3547  Map of search patterns to permission masks. If a directory name matches  =cut
 one of the patterns, that directory and all its members and subdirectories  
 will be assigned the new pattern. For example, the following would  
 assign 01664 to most files, but would use 01777 for directories named C<tmp>.  
3548    
3549      Tracer::SetPermissions($dirName, 'fig', 01664, '^tmp$' => 01777);  sub CombineURL {
3550        # Get the parameters.
3551        my ($text, $url) = @_;
3552        # Declare the return variable.
3553        my $retVal = CGI::escapeHTML($text);
3554        # Verify the URL.
3555        if (defined($url) && $url =~ m!http://!i) {
3556            # It's good, so we apply it to the text.
3557            $retVal = "<a href=\"$url\">$retVal</a>";