texlive[54167] Master/tlpkg: TLPSRC.pm (from_file),

[commits+karl at tug.org]

Revision: 54167
          http://tug.org/svn/texlive?view=revision&revision=54167
Author:   karl
Date:     2020-03-07 23:55:02 +0100 (Sat, 07 Mar 2020)
Log Message:
-----------
TLPSRC.pm (from_file),
TeXCatalogue.pm (beautify): collapse multiple whitespace.
(Still on the quest for identical longdesc results from our various
sources.)

Modified Paths:
--------------
    trunk/Master/tlpkg/TeXLive/TLPSRC.pm
    trunk/Master/tlpkg/TeXLive/TeXCatalogue.pm
    trunk/Master/tlpkg/doc/releng.txt

Modified: trunk/Master/tlpkg/TeXLive/TLPSRC.pm
===================================================================
--- trunk/Master/tlpkg/TeXLive/TLPSRC.pm	2020-03-07 22:52:33 UTC (rev 54166)
+++ trunk/Master/tlpkg/TeXLive/TLPSRC.pm	2020-03-07 22:55:02 UTC (rev 54167)
@@ -232,14 +232,14 @@
     die "Cannot deduce name from file argument and name tag not found";
   }
   #
-  # We should probably call TeXCatalogue::beautify(), but trailing
-  # whitespace seems to be the only thing that comes up in practice. We
-  # want the parsing from .tlpsrc to result in exactly the same string,
-  # including spaces, as parsing from texlive.tlpdb. Otherwise
-  # tl-update-tlpdb's tlpdb_catalogue_compare will think the strings
-  # are always different.
-  $shortdesc =~ s/\s+$//g;
-  $longdesc =~ s/\s+$//g;
+  # We should call TeXCatalogue::beautify(), but let's be lazy since not
+  # everything comes up in practice. We want the parsing from .tlpsrc to
+  # result in exactly the same string, including spaces, as parsing from
+  # texlive.tlpdb. Otherwise tl-update-tlpdb's tlpdb_catalogue_compare
+  # will think the strings are always different.
+  $shortdesc =~ s/\s+$//g;  # rm trailing whitespace (shortdesc)
+  $longdesc =~ s/\s+$//g;   # rm trailing whitespace (longdesc)
+  $longdesc =~ s/\s\s+/ /g; # collapse multiple whitespace characters to one
   #
   $self->name($name);
   $self->category($category);

Modified: trunk/Master/tlpkg/TeXLive/TeXCatalogue.pm
===================================================================
--- trunk/Master/tlpkg/TeXLive/TeXCatalogue.pm	2020-03-07 22:52:33 UTC (rev 54166)
+++ trunk/Master/tlpkg/TeXLive/TeXCatalogue.pm	2020-03-07 22:55:02 UTC (rev 54167)
@@ -140,7 +140,8 @@
   $txt =~ s/\n/ /g;  # make one line
   $txt =~ s/^\s+//g; # rm leading whitespace
   $txt =~ s/\s+$//g; # rm trailing whitespace
-  $txt =~ s/\s\s+/ /g; # multiple spaces to one
+  $txt =~ s/\s\s+/ /g; # collapse multiple whitespace characters to one
+  $txt =~ s/\t/ /g;    # tabs to spaces
   # transliterate to ascii: it allows the final tlpdb to be pure ascii,
   # avoiding problems since we don't control the user's terminal encoding
   return Text::Unidecode::unidecode($txt);

Modified: trunk/Master/tlpkg/doc/releng.txt
===================================================================
--- trunk/Master/tlpkg/doc/releng.txt	2020-03-07 22:52:33 UTC (rev 54166)
+++ trunk/Master/tlpkg/doc/releng.txt	2020-03-07 22:55:02 UTC (rev 54167)
@@ -164,7 +164,7 @@
   unzip -q newperl.zip  # replace All files
   comm -13 /tmp/now /tmp/new >/tmp/c2
   comm -23 /tmp/now /tmp/new >/tmp/c1
-  wc -l /tmp/c1 /tmp/c2 # should not be everything
+  wc -l /tmp/c1 /tmp/c2 # should be about 100-200, not 1500+
   svn rm `cat /tmp/c1`
   svn add --parents --no-auto-props `cat /tmp/c2`
   svn status | fgrep \? # shouldn't be any besides newperl.zip