branch master updated: texi2any --html performance improvement - XS override

[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 3a1049f252 texi2any --html performance improvement - XS override
3a1049f252 is described below

commit 3a1049f252eb2778cd91560a287bfa6ff6219c91
Author: Gavin Smith <gavinsmith0123@gmail.com>
AuthorDate: Sat Jun 4 01:32:21 2022 +0100

    texi2any --html performance improvement - XS override
    
    * tp/Texinfo/XS/MiscXS.xs,
    * tp/Texinfo/XS/misc.c (xs_xml_protect_text): New function.
    * tp/Texinfo/Convert/Converter.pm: Override xml_protect_text
    with XS implementation.
---
 ChangeLog                       |  9 +++++++
 tp/Texinfo/Convert/Converter.pm | 15 +++++++++++-
 tp/Texinfo/XS/MiscXS.xs         | 27 +++++++++++++++++++--
 tp/Texinfo/XS/misc.c            | 52 ++++++++++++++++++++++++++++++++++++++++-
 tp/Texinfo/XS/miscxs.h          |  1 +
 5 files changed, 100 insertions(+), 4 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 1e14b62b80..0397720527 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,12 @@
+2022-06-05  Gavin Smith  <gavinsmith0123@gmail.com>
+
+       texi2any --html performance improvement - XS override
+
+       * tp/Texinfo/XS/MiscXS.xs,
+       * tp/Texinfo/XS/misc.c (xs_xml_protect_text): New function.
+       * tp/Texinfo/Convert/Converter.pm: Override xml_protect_text
+       with XS implementation.
+
 2022-06-05  Gavin Smith  <gavinsmith0123@gmail.com>
 
        texi2any --html performance improvement
diff --git a/tp/Texinfo/Convert/Converter.pm b/tp/Texinfo/Convert/Converter.pm
index 46664be890..c4e8528e63 100644
--- a/tp/Texinfo/Convert/Converter.pm
+++ b/tp/Texinfo/Convert/Converter.pm
@@ -57,6 +57,18 @@ xml_accents
 
 $VERSION = '6.8dev';
 
+our $module_loaded = 0;
+sub import {
+  if (!$module_loaded) {
+    Texinfo::XSLoader::override(
+      "Texinfo::Convert::Converter::xml_protect_text",
+      "Texinfo::MiscXS::xml_protect_text");
+    $module_loaded = 1;
+  }
+  # The usual import method
+  goto &Exporter::import;
+}
+
 
 my %defaults = (
   'documentlanguage'     => undef,
@@ -1383,7 +1395,8 @@ sub xml_format_text_with_numeric_entities($$)
   return $text;
 }
 
-sub xml_protect_text($$)
+# Note: has an XS override
+sub xml_protect_text
 {
   my $self = shift;
   my $text = shift;
diff --git a/tp/Texinfo/XS/MiscXS.xs b/tp/Texinfo/XS/MiscXS.xs
index 04e9454c5c..cd1307ff48 100644
--- a/tp/Texinfo/XS/MiscXS.xs
+++ b/tp/Texinfo/XS/MiscXS.xs
@@ -13,7 +13,7 @@
 
 MODULE = Texinfo::MiscXS  PACKAGE = Texinfo::MiscXS  PREFIX = xs_
 
-#  Copyright 2016 Free Software Foundation, Inc.
+#  Copyright 2016-2022 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
@@ -86,7 +86,7 @@ xs_unicode_text (text_in, ...)
          if (SvOK(ST(1)))
            in_code = (int) SvIV(ST(1));
        }
-     /* Make sure the input is in UTF8. */
+     /* Make sure the input is in UTF-8. */
      if (!SvUTF8 (text_in))
        sv_utf8_upgrade (text_in);
 
@@ -132,3 +132,26 @@ xs_parse_texi_regex (text)
      PUSHs(sv_newmortal());
      sv_setpv((SV*)ST(5), new_text);
      SvUTF8_on(ST(5));
+
+SV *
+xs_xml_protect_text (self, text_in)
+     SV *self
+     SV *text_in
+ PREINIT:
+     char *text;
+     char *retval;
+ CODE:
+     /* Make sure the input is in UTF-8. */
+     if (!SvUTF8 (text_in))
+       sv_utf8_upgrade (text_in);
+
+     text = SvPV_nolen (text_in);
+
+     retval = xs_xml_protect_text (text);
+
+     RETVAL = newSVpv (retval, 0);
+     SvUTF8_on (RETVAL);
+
+ OUTPUT:
+     RETVAL
+
diff --git a/tp/Texinfo/XS/misc.c b/tp/Texinfo/XS/misc.c
index 91b4708fdd..c3b3427043 100644
--- a/tp/Texinfo/XS/misc.c
+++ b/tp/Texinfo/XS/misc.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
@@ -634,3 +634,53 @@ void xs_parse_texi_regex (SV *text_in,
 
   return;
 }
+
+char *
+xs_xml_protect_text (char *text)
+{
+  char *p, *q;
+  static char *new;
+  int new_space, new_len;
+
+  dTHX; /* Perl boilerplate. */
+
+  p = text;
+  new_space = strlen (text);
+  new = realloc (new, new_space + 1);
+  new_len = 0;
+
+#define ADDN(s, n) \
+  if (new_len + n - 1 >= new_space - 1)           \
+    {                                             \
+      new_space += n;                             \
+      new = realloc (new, (new_space *= 2) + 1);  \
+    }                                             \
+  memcpy(new + new_len, s, n);                    \
+  new_len += n;
+
+  while (1)
+    {
+      q = p + strcspn (p, "<>&\"");
+      ADDN(p, q - p);
+      if (!*q)
+        break;
+      switch (*q)
+        {
+        case '<':
+          ADDN("&lt;", 4);
+          break;
+        case '>':
+          ADDN("&gt;", 4);
+          break;
+        case '&':
+          ADDN("&amp;", 5);
+          break;
+        case '"':
+          ADDN("&quot;", 6);
+          break;
+        }
+      p = q + 1;
+    }
+  new[new_len] = '\0';
+  return new;
+}
diff --git a/tp/Texinfo/XS/miscxs.h b/tp/Texinfo/XS/miscxs.h
index 745e6ec8e9..4eca89ccf3 100644
--- a/tp/Texinfo/XS/miscxs.h
+++ b/tp/Texinfo/XS/miscxs.h
@@ -9,3 +9,4 @@ void xs_parse_texi_regex (SV *text,
                           char **,
                           char **,
                           char **);
+char *xs_xml_protect_text (char *);