branch master updated: Remove old TODO

[Gavin D. Smith]

This is an automated email from the git hooks/post-receive script.

gavin pushed a commit to branch master
in repository texinfo.

The following commit(s) were added to refs/heads/master by this push:
     new 516b8b5cfb Remove old TODO
516b8b5cfb is described below

commit 516b8b5cfb364628d1e284770d9f9a19fc0658b1
Author: Gavin Smith <gavinsmith0123@gmail.com>
AuthorDate: Sun Jan 9 17:13:20 2022 +0000

    Remove old TODO
    
    * tp/Texinfo/XS/parsetexi/input.c,
    * tp/Texinfo/XS/parsetexi/macro.c: Do not define _GNU_SOURCE as
    gnulib should deal with this.
    * tp/Texinfo/XS/parsetexi/api.c (element_to_perl_hash): Add a
    comment regarding use of newSVpv.
    
    * tp/Texinfo/XS/parsetexi/TODO: Remove obsolete file.  Report from
    Patrice.
---
 ChangeLog                       |  13 +++
 tp/Texinfo/XS/parsetexi/TODO    | 204 ----------------------------------------
 tp/Texinfo/XS/parsetexi/api.c   |   5 +-
 tp/Texinfo/XS/parsetexi/input.c |   4 +-
 tp/Texinfo/XS/parsetexi/macro.c |   4 +-
 5 files changed, 18 insertions(+), 212 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 2a07409075..bf0343ceae 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,16 @@
+2022-01-09  Gavin Smith  <gavinsmith0123@gmail.com>
+
+       Remove old TODO
+
+       * tp/Texinfo/XS/parsetexi/input.c,
+       * tp/Texinfo/XS/parsetexi/macro.c: Do not define _GNU_SOURCE as
+       gnulib should deal with this.
+       * tp/Texinfo/XS/parsetexi/api.c (element_to_perl_hash): Add a
+       comment regarding use of newSVpv.
+
+       * tp/Texinfo/XS/parsetexi/TODO: Remove obsolete file.  Report from 
+       Patrice.
+
 2022-01-09  Patrice Dumas  <pertusus@free.fr>
 
        * tp/Texinfo/ParserNonXS.pm (parse_texi_line),
diff --git a/tp/Texinfo/XS/parsetexi/TODO b/tp/Texinfo/XS/parsetexi/TODO
deleted file mode 100644
index 7d210e3ba1..0000000000
--- a/tp/Texinfo/XS/parsetexi/TODO
+++ /dev/null
@@ -1,204 +0,0 @@
-* conditionals on @set/@value, output format
-* implement command_warn properly
-* floats
-* Check everywhere we store values, we check properly what to do if the
-  values are already defined.
-* check valid nestings
-* Implement more parser options
-* Character encoding issues - read and pass to Perl correctly.
-* gnulib integration for strrchr, strchrnul, asprintf, etc.  Eliminate
-  #define _GNU_SOURCE.
-* gettext integration
-* figure out what to do with parse_texi_line, etc.  Integration with the 
-test suites.
-* find where "strlen" and "text_append" are used and try to remove them, for 
-efficiency.  Also newSVpv (..., 0);
-* Add "TODO" anywhere in the code with explanations of what is not done yet.
-
-================================================================
-
-Texinfo::Report::gdt and parse_texi_line -
-
-Multiple Texinfo::Parser objects are created in a run of texi2any.  The
-main one is to parse the input file, but it is also used by gdt - the 
-string translation function.  (I haven't ruled out that it is also used 
-somewhere else as well.)
-
-This makes replacing the Texinfo::Parser module more complicated.
-
-gdt parses pieces Texinfo code that are the translations of strings in 
-po_document/texinfo_document.pot.  It does this by adding "@value" 
-before strings surrounded by braces, and then calling "parse_texi_line" 
-with these @value's set to the appropriate value.  These values could be 
-a string, a hash reference (representing a tree element), or an array 
-reference (representing a "content array" of tree elements).
-
-parse_texi_line returns a Perl Texinfo tree which is at last converted 
-into the output Info/HTML/etc.
-
-These values that are not strings are a challenge for how they can be 
-passed back into C.  We have code (in api.c) to translate our C 
-representations of the Texinfo tree to Perl data structures, but not 
-vice versa.
-
-We deal with this by leaving the @value's undefined, letting an element 
-representing a use of an unused @value appear in the parse tree, and 
-then substituting in the parse trees for the @value's afterwards, in 
-Texinfo::Report::gdt.  See diff below.
-
-Other suggestions:
-* Don't call parse_texi_line at all.  This would be a simplification of 
-the overall structure of the makeinfo program and might not lose much 
-flexibility in return.  Most of the strings in texinfo_document.pot 
-don't use any Texinfo commands, and the others could be expressed 
-directly in the output format, e.g. instead of
-
-msgid "{category} on {class}: @strong{{name}}
-
-it could be
-
-msgid "{category} on {class}: <strong>{name}</strong>"
-
-Some of the translation strings use Texinfo commands for characters, 
-e.g. "@'e".  This would have to be replaced with whatever method other 
-programs using gettext use for special characters.  It would also stop 
-translation strings being reused across output formats, creating more 
-work for translators.
-
-* Write code translating Perl tree elements into C data structures and 
-allow them to be passed to the parser as @value's.  I am very reluctant 
-to do this.
-
-To use the current implementation, do the following to 
-tp/Texinfo/Report.pm:
-
-Index: Texinfo/Report.pm
-===================================================================
---- Texinfo/Report.pm   (revision 6344)
-+++ Texinfo/Report.pm   (working copy)
-@@ -49,6 +49,7 @@ use File::Basename;
- use Locale::Messages;
- # to be able to load a parser if none was given to gdt.
- use Texinfo::Parser;
-+use Parsetexi;
- 
- # return the errors and warnings
- sub errors($)
-@@ -244,7 +245,8 @@ sub _encode_i18n_string($$)
-   return Encode::decode($encoding, $string);
- }
- 
--# handle translations of in-document strings.
-+# Get document translation - handle translations of in-document strings.
-+# Return a parsed Texinfo tree
- sub gdt($$;$$)
- {
-   my $self = shift;
-@@ -376,7 +378,11 @@ sub gdt($$;$$)
-     $translation_result =~ s/\{($re)\}/\@value\{_$1\}/g;
-     foreach my $substitution(keys %$context) {
-       #print STDERR "$translation_result $substitution 
$context->{$substitution}\n";
--      $parser_conf->{'values'}->{'_'.$substitution} = 
$context->{$substitution};
-+      if (!ref($context->{$substitution})) {
-+        #warn "setting $substitution to $context->{$substitution}\n";
-+        $parser_conf->{'values'}->{'_'.$substitution}
-+          = $context->{$substitution};
-+      }
-     }
-   }
- 
-@@ -399,7 +405,8 @@ sub gdt($$;$$)
-     }
-   }
-   #my $parser = Texinfo::Parser::parser($parser_conf);
--  my $parser = Texinfo::Parser::simple_parser($parser_conf);
-+  #my $parser = Texinfo::Parser::simple_parser($parser_conf);
-+  my $parser = Parsetexi::parser($parser_conf);
-   if ($parser->{'DEBUG'}) {
-     print STDERR "GDT $translation_result\n";
-   }
-@@ -411,10 +418,67 @@ sub gdt($$;$$)
-   } else {
-     $tree = $parser->parse_texi_line($translation_result);
-   }
-+  #warn '--------------------------------------', "\n";
-+  #warn Texinfo::Parser::_print_tree ($tree);
-+  $tree = _substitute ($tree, $context);
-+  #warn Texinfo::Parser::_print_tree ($tree);
-+  #warn '--------------------------------------', "\n";
-   return $tree;
- }
- 
-+sub _substitute_element_array ($$);
-+sub _substitute_element_array ($$) {
-+  my $array = shift; my $context = shift;
- 
-+  # "the push @{$current->{'contents'}}, {}; prevents a trailing
-+  # text to be merged, to avoid having the value tree modified."
-+
-+  # It's not necessary now, but we are yet to update the test
-+  # results to remove the extra {}'s.
-+
-+  @{$array} = map {
-+    if ($_->{'cmdname'} and $_->{'cmdname'} eq 'value'
-+      and $_->{'type'}
-+    ) {
-+      my $name = $_->{'type'};
-+      $name =~ s/^_//;
-+      if (ref($context->{$name}) eq 'HASH') {
-+        ( $context->{$name} , {} );
-+      } elsif (ref($context->{$name}) eq 'ARRAY') {
-+        ( @{$context->{$name}} , {} );
-+      }
-+    } else {
-+      _substitute($_, $context);
-+      ( $_ );
-+    }
-+  } @{$array};
-+}
-+
-+sub _substitute ($$);
-+sub _substitute ($$) {
-+  my $tree = shift; my $context = shift;
-+
-+  if ($tree->{'contents'}) {
-+    _substitute_element_array ($tree->{'contents'}, $context);
-+  }
-+
-+  if ($tree->{'args'}) {
-+    _substitute_element_array ($tree->{'args'}, $context);
-+  }
-+
-+  # Used for @email and @url
-+  if ($tree->{'extra'} and $tree->{'extra'}{'brace_command_contents'}) {
-+    for my $arg (@{$tree->{'extra'}{'brace_command_contents'}}) {
-+      if ($arg) {
-+        _substitute_element_array ($arg, $context);
-+      }
-+    }
-+  }
-+
-+  return $tree;
-+}
-+
-+
- 1;
- 
- __END__
-
-
-===================================================================
-
-
-Integration with rest of Perl code:
-
-Passing data as Perl code to be evaluated in very slow, which limits the 
-speed-up that is obtained. One way round this would be to access the 
-parse tree in the Perl code through an API. This would be done in two 
-stages:
-
-* Convert makeinfo backends to use a Perl stub for the C API. The C 
-parser will not be used at this point.
-* Substitute the XS API implementation for the Perl API when the C 
-parser is good enough.
-
diff --git a/tp/Texinfo/XS/parsetexi/api.c b/tp/Texinfo/XS/parsetexi/api.c
index 22f974b366..59d4ded019 100644
--- a/tp/Texinfo/XS/parsetexi/api.c
+++ b/tp/Texinfo/XS/parsetexi/api.c
@@ -1,4 +1,4 @@
-/* Copyright 2010-2021 Free Software Foundation, Inc.
+/* Copyright 2010-2022 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -294,7 +294,8 @@ element_to_perl_hash (ELEMENT *e)
       sv = newSVpv (command_name(e->cmd), 0);
       hv_store (e->hv, "cmdname", strlen ("cmdname"), sv, 0);
 
-      /* TODO: Same optimizations as for 'type'. */
+      /* Note we could optimize the call to newSVpv here and
+         elsewhere by passing an appropriate second argument. */
     }
 
   /* A lot of special cases for when an empty contents array should be
diff --git a/tp/Texinfo/XS/parsetexi/input.c b/tp/Texinfo/XS/parsetexi/input.c
index 990ad84f38..81120f81a0 100644
--- a/tp/Texinfo/XS/parsetexi/input.c
+++ b/tp/Texinfo/XS/parsetexi/input.c
@@ -1,4 +1,4 @@
-/* Copyright 2010-2020 Free Software Foundation, Inc.
+/* Copyright 2010-2022 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -13,8 +13,6 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>. */
 
-#define _GNU_SOURCE
-
 #include <config.h>
 
 #include <stdlib.h>
diff --git a/tp/Texinfo/XS/parsetexi/macro.c b/tp/Texinfo/XS/parsetexi/macro.c
index 06e3ed4826..36bb41d9d9 100644
--- a/tp/Texinfo/XS/parsetexi/macro.c
+++ b/tp/Texinfo/XS/parsetexi/macro.c
@@ -1,4 +1,4 @@
-/* Copyright 2010-2019 Free Software Foundation, Inc.
+/* Copyright 2010-2022 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -13,8 +13,6 @@
    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>. */
 
-#define _GNU_SOURCE
-
 #include <config.h>
 #include <stdlib.h>
 #include <string.h>