How can I split the log file depending on the input file?

Question

I have a large document that I have split into several separate documents. I have one main.tex file and several file_i.tex files which are included into main.tex using the \input-command. It also happens that this including is nested, i.e. that one of file_i.tex files included another file.

The log file main.log that is generated looks something like this:

blabla
(main.tex
blabla
blabla
(file_1.tex
blabla
)
(file_2.tex
blablabla
(file_2_1.tex)
)
blabla
)

I wonder if there exists some way/script that splits this file into e.g. part_main.log, part_file_1.log, part_file_2.log and part_file_2_1.log.

It doesn't seem to difficult to write an ad-hoc script to do this, but it also seems such a common problem, that I would expect some well-tested scripts to already be available.

Answer 1

I don't know of an "official" script but I wrote a little perl script some time ago for a similar task. It is kind of a dirty hack and sparsely tested, but it might be useful for others as well. Improvements are welcome.

The script expects TeX's logfile as a parameter and generates separate .log files for each included module.

#!/usr/bin/perl

use File::Basename;

die("Syntax: splitlog <logfile>") unless ($#ARGV == 1);

@log = <ARGV>;
push(@files, $ARGV[0]);
foreach (@log) {
   chomp;
   while (length($_) > 0) {
      if (/^\(([^)]+)$/) {
         push(@files, $1);
         $_ = "";
         $fname = basename($1);
         open(LOG, ">$fname.log");
      }
      elsif (/^(\)+)\s*(.*)$/) {
         for ($i=0; $i < length($1); ++$i) {
            close(LOG);
            pop(@files);
         }
         if ($#files > 0) {
            $fname = basename($files[-1]);
            open(LOG, ">>$fname.log");
         }
         $_ = $2;
      }
      else {
         print LOG "$_\n";
         $_ = "";
      }
   }
}