[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
branch master updated: * tp/Texinfo/Structuring.pm (setup_index_entry_ke
|
From: |
Patrice Dumas |
|
Subject: |
branch master updated: * tp/Texinfo/Structuring.pm (setup_index_entry_keys_formatting): remove ascii_punctuation obsolete option set. |
|
Date: |
Sun, 07 Jan 2024 18:36:24 -0500 |
This is an automated email from the git hooks/post-receive script.
pertusus pushed a commit to branch master
in repository texinfo.
The following commit(s) were added to refs/heads/master by this push:
new b4296ae27c * tp/Texinfo/Structuring.pm
(setup_index_entry_keys_formatting): remove ascii_punctuation obsolete option
set.
b4296ae27c is described below
commit b4296ae27c7409b45d7af82e736df45e295120f3
Author: Patrice Dumas <pertusus@free.fr>
AuthorDate: Mon Jan 8 00:36:23 2024 +0100
* tp/Texinfo/Structuring.pm (setup_index_entry_keys_formatting):
remove ascii_punctuation obsolete option set.
* tp/Texinfo/Structuring.pm (index_entry_element_sort_string)
(_index_entry_element_sort_string_key)
(setup_sortable_index_entries), tp/t/test_sort.t, tp/t/test_utils.pl:
add an argument with document information for XS for
index_entry_element_sort_string and
_index_entry_element_sort_string_key to retrieve the document
descriptor in XS. Use setup_sortable_index_entries
$customization_information as a source of that information. If it is
a converter based on Texinfo::Convert::Converter, the information is
there, if not, it needs to be added explcitely. Add the information
explicitely setting 'document_descriptor' to
document->document_descriptor() in test_sort.t and test_utils.pl.
* tp/Texinfo/XS/convert/indices_in_conversion.c
(index_entry_element_sort_string): implement in C.
* tp/Texinfo/XS/main/get_perl_info.c (find_index_entry_sv): add based
on get_sv_index_entries_sorted_by_letter code.
* tp/Texinfo/XS/main/get_perl_info.c
(copy_sv_options_for_convert_text): get 'code' from perl.
* tp/Texinfo/XS/main/get_perl_info.c (find_index_entry_subentry)
(subentry_hv_parent, find_subentry_index_command_sv)
(find_element_from_sv): find subentry C element based on perl element,
by finding the 'subentry parent' index entry C element going down
subentry levels, and then find the C subentry element going up
subentry levels.
* tp/Texinfo/XS/main/get_perl_info.c
(find_element_extra_index_entry_sv)
(find_index_entry_associated_hv, find_element_from_sv): have
find_element_extra_index_entry_sv return an index entry, such that the
caller can find the element based on entry_associated_element or
entry_element. Add find_index_entry_associated_hv to go through both
entry_associated_element and entry_element to find the C element
associated to a perl element. Use that code in find_element_from_sv.
* tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
(index_entry_element_sort_string): XS interface for
index_entry_element_sort_string. Not used as it is slower than the
perl...
* tp/Texinfo/XS/main/get_perl_info.c (debug_print_element_hv)
(debug_print_element_sv): add debug functions that print information
on perl tree elements.
---
ChangeLog | 52 ++++
tp/TODO | 12 -
tp/Texinfo/Convert/LaTeX.pm | 3 +-
tp/Texinfo/Structuring.pm | 42 ++-
tp/Texinfo/XS/Makefile.am | 4 +-
tp/Texinfo/XS/convert/indices_in_conversion.c | 55 ++++
tp/Texinfo/XS/convert/indices_in_conversion.h | 6 +-
tp/Texinfo/XS/main/debug.c | 3 +-
tp/Texinfo/XS/main/get_perl_info.c | 326 +++++++++++++++++----
tp/Texinfo/XS/main/get_perl_info.h | 4 +
.../XS/structuring_transfo/StructuringTransfoXS.xs | 34 +++
tp/t/test_sort.t | 4 +
tp/t/test_utils.pl | 2 +
13 files changed, 460 insertions(+), 87 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index ac9dd5b8ad..417ffd4736 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,55 @@
+2024-01-07 Patrice Dumas <pertusus@free.fr>
+
+ * tp/Texinfo/Structuring.pm (setup_index_entry_keys_formatting):
+ remove ascii_punctuation obsolete option set.
+
+ * tp/Texinfo/Structuring.pm (index_entry_element_sort_string)
+ (_index_entry_element_sort_string_key)
+ (setup_sortable_index_entries), tp/t/test_sort.t, tp/t/test_utils.pl:
+ add an argument with document information for XS for
+ index_entry_element_sort_string and
+ _index_entry_element_sort_string_key to retrieve the document
+ descriptor in XS. Use setup_sortable_index_entries
+ $customization_information as a source of that information. If it is
+ a converter based on Texinfo::Convert::Converter, the information is
+ there, if not, it needs to be added explcitely. Add the information
+ explicitely setting 'document_descriptor' to
+ document->document_descriptor() in test_sort.t and test_utils.pl.
+
+ * tp/Texinfo/XS/convert/indices_in_conversion.c
+ (index_entry_element_sort_string): implement in C.
+
+ * tp/Texinfo/XS/main/get_perl_info.c (find_index_entry_sv): add based
+ on get_sv_index_entries_sorted_by_letter code.
+
+ * tp/Texinfo/XS/main/get_perl_info.c
+ (copy_sv_options_for_convert_text): get 'code' from perl.
+
+ * tp/Texinfo/XS/main/get_perl_info.c (find_index_entry_subentry)
+ (subentry_hv_parent, find_subentry_index_command_sv)
+ (find_element_from_sv): find subentry C element based on perl element,
+ by finding the 'subentry parent' index entry C element going down
+ subentry levels, and then find the C subentry element going up
+ subentry levels.
+
+ * tp/Texinfo/XS/main/get_perl_info.c
+ (find_element_extra_index_entry_sv)
+ (find_index_entry_associated_hv, find_element_from_sv): have
+ find_element_extra_index_entry_sv return an index entry, such that the
+ caller can find the element based on entry_associated_element or
+ entry_element. Add find_index_entry_associated_hv to go through both
+ entry_associated_element and entry_element to find the C element
+ associated to a perl element. Use that code in find_element_from_sv.
+
+ * tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
+ (index_entry_element_sort_string): XS interface for
+ index_entry_element_sort_string. Not used as it is slower than the
+ perl...
+
+ * tp/Texinfo/XS/main/get_perl_info.c (debug_print_element_hv)
+ (debug_print_element_sv): add debug functions that print information
+ on perl tree elements.
+
2024-01-07 Patrice Dumas <pertusus@free.fr>
* tp/Texinfo/XS/convert/convert_html.c (convert_printindex_command):
diff --git a/tp/TODO b/tp/TODO
index 5db2a0c266..179ba95fe8 100644
--- a/tp/TODO
+++ b/tp/TODO
@@ -20,18 +20,6 @@ Document that Texinfo::Document::rebuild_document or
Texinfo::Document::rebuild_tree should be called after tree modifications
if the parser is XS but converters are perl.
-Find perl index element in C:
-$element->{'extra'}->{'index_entry'} = [$index_name, $number];
-in C:
-index[$index_name][number] = { 'index_name' => $index_name,
- 'entry_element' => $element,
- 'entry_number' => $number,
- };
-(find index j based on index_name, could be a bsearch or a lin search)
- INDEX_ENTRY *index_entry;
- index_entry = idx->index_entries[j];
- main_entry_element = index_entry->entry_element;
-
Bugs
====
diff --git a/tp/Texinfo/Convert/LaTeX.pm b/tp/Texinfo/Convert/LaTeX.pm
index 03cd19e2e7..e178016a0e 100644
--- a/tp/Texinfo/Convert/LaTeX.pm
+++ b/tp/Texinfo/Convert/LaTeX.pm
@@ -2454,7 +2454,8 @@ sub _index_entry($$)
# always setup a string to sort with as we may use commands
my $convert_to_text_options = {%$options, 'code' => $in_code};
my $sort_string
- = Texinfo::Structuring::index_entry_element_sort_string($entry,
+ = Texinfo::Structuring::index_entry_element_sort_string(
+ $self, $entry,
$subindex_command,
$convert_to_text_options, 1);
my $result = '';
diff --git a/tp/Texinfo/Structuring.pm b/tp/Texinfo/Structuring.pm
index 89baa4ee2a..abbb3fb4e8 100644
--- a/tp/Texinfo/Structuring.pm
+++ b/tp/Texinfo/Structuring.pm
@@ -114,8 +114,11 @@ our %XS_overrides = (
"Texinfo::Structuring::_XS_unsplit"
=> "Texinfo::StructTransfXS::unsplit",
-# "Texinfo::Structuring::index_entry_element_sort_string"
-# => "Texinfo::StructTransfXS::index_entry_element_sort_string",
+ # TODO the XS override is slower than the perl function.
+ # One possible reason could be that the text options are read
+ # from perl for each entry instead of once for each index.
+ #"Texinfo::Structuring::index_entry_element_sort_string"
+ # => "Texinfo::StructTransfXS::index_entry_element_sort_string",
# Not useful for HTML as functions, as the calling functions are
# already overriden
@@ -2270,7 +2273,7 @@ sub setup_index_entry_keys_formatting($)
{
my $customization_info = shift;
- my $options = {'ascii_punctuation' => 1,
+ my $options = {
Texinfo::Convert::Text::copy_options_for_convert_text(
$customization_info)};
if (not $customization_info->get_conf('ENABLE_ENCODING')
@@ -2281,9 +2284,11 @@ sub setup_index_entry_keys_formatting($)
return $options;
}
-# can be used for subentries
-sub index_entry_element_sort_string($$$;$)
+# can be used for subentries.
+# $DOCUMENT_INFO is used in XS to retrieve the document.
+sub index_entry_element_sort_string($$$$;$)
{
+ my $document_info = shift;
my $main_entry = shift;
my $index_entry_element = shift;
my $options = shift;
@@ -2312,16 +2317,17 @@ sub index_entry_element_sort_string($$$;$)
return $sort_string;
}
-sub _index_entry_element_sort_string_key($$$$;$)
+sub _index_entry_element_sort_string_key($$$$$;$)
{
+ my $document_info = shift;
my $main_entry = shift;
my $index_entry_element = shift;
my $options = shift;
my $collator = shift;
my $prefer_reference_element = shift;
- my $sort_string = index_entry_element_sort_string ($main_entry,
- $index_entry_element,
+ my $sort_string = index_entry_element_sort_string ($document_info,
+ $main_entry, $index_entry_element,
$options, $prefer_reference_element);
# This avoids varying results depending on whether the string is
@@ -2397,6 +2403,12 @@ sub _converter_or_registrar_line_warn($$$$)
}
}
+# There is no neeed for document information in Perl, however, in XS
+# it is needed to retrieve the Tree elements in the C structures.
+# $CUSTOMIZATION_INFORMATION is used as the source of document
+# information. It should already be set if it is a converter based
+# on Texinfo::Convert::Converter, but otherwise it should be set by
+# the caller, setting 'document_descriptor' to document->document_descriptor().
sub setup_sortable_index_entries($$$$$;$)
{
my $registrar = shift;
@@ -2469,8 +2481,8 @@ sub setup_sortable_index_entries($$$$$;$)
my $convert_to_text_options = {%$options,
'code' => $indices_information->{$entry_index_name}->{'in_code'}};
my ($entry_key, $sort_entry_key)
- = _index_entry_element_sort_string_key($index_entry,
- $main_entry_element,
+ = _index_entry_element_sort_string_key($customization_information,
+ $index_entry, $main_entry_element,
$convert_to_text_options, $entries_collator);
my @entry_keys;
my @sort_entry_keys;
@@ -2498,8 +2510,8 @@ sub setup_sortable_index_entries($$$$$;$)
$subentry_nr ++;
$subentry = $subentry->{'extra'}->{'subentry'};
my ($subentry_key, $sort_subentry_key)
- = _index_entry_element_sort_string_key($index_entry,
- $subentry, $convert_to_text_options,
+ =
_index_entry_element_sort_string_key($customization_information,
+ $index_entry, $subentry, $convert_to_text_options,
$entries_collator);
if ($subentry_key !~ /\S/) {
my $entry_cmdname = $main_entry_element->{'cmdname'};
@@ -2879,10 +2891,14 @@ I<$node> is a node tree element. Find the node
I<$node> children based
on the sectioning structure. For the node associated with C<@top>
sectioning command, the sections associated with parts are considered.
-=item $sort_string = index_entry_element_sort_string($main_entry,
$index_entry_element, $options, $prefer_reference_element)
+=item $sort_string = index_entry_element_sort_string($document_info,
$main_entry, $index_entry_element, $options, $prefer_reference_element)
X<C<index_entry_element_sort_string>>
Return a string suitable as a sort string, for index entries.
+I<$document_info> is used by C code to retrieve the document data,
+using the C<document_descriptor> key. I<$document_info> can be a
+converter based on L<Texinfo::Convert::Converter>, otherwise
+C<document_descriptor> need, in general, to be set up explicitely.
The tree element index entry processed is I<$index_entry_element>,
and can be a C<@subentry>. I<$main_entry> is the main index entry
that can be used to gather information. The I<$options> are options
diff --git a/tp/Texinfo/XS/Makefile.am b/tp/Texinfo/XS/Makefile.am
index 7658a51b6c..3d85482b84 100644
--- a/tp/Texinfo/XS/Makefile.am
+++ b/tp/Texinfo/XS/Makefile.am
@@ -350,13 +350,15 @@ nodist_StructuringTransfoXS_la_SOURCES = \
CLEANFILES += \
structuring_transfo/StructuringTransfoXS.c
StructuringTransfoXS_la_SOURCES = \
+ convert/indices_in_conversion.h \
+ convert/indices_in_conversion.c \
structuring_transfo/transformations.c \
structuring_transfo/transformations.h
EXTRA_DIST += structuring_transfo/StructuringTransfoXS.xs
# locate include files under out-of-source builds.
-StructuringTransfoXS_la_CPPFLAGS = -I$(srcdir)/main
-I$(srcdir)/structuring_transfo $(AM_CPPFLAGS) $(GNULIB_CPPFLAGS)
$(XSLIBS_CPPFLAGS)
+StructuringTransfoXS_la_CPPFLAGS = -I$(srcdir)/main
-I$(srcdir)/structuring_transfo -I$(srcdir)/convert $(AM_CPPFLAGS)
$(GNULIB_CPPFLAGS) $(XSLIBS_CPPFLAGS)
StructuringTransfoXS_la_CFLAGS = $(XSLIBS_CFLAGS)
StructuringTransfoXS_la_LIBADD = libtexinfoxs.la libtexinfo.la
$(top_builddir)/gnulib/lib/libgnu.la
StructuringTransfoXS_la_LDFLAGS = $(XSLIBS_LDFLAGS) $(LTLIBICONV)
$(LTLIBUNISTRING)
diff --git a/tp/Texinfo/XS/convert/indices_in_conversion.c
b/tp/Texinfo/XS/convert/indices_in_conversion.c
index 3774dec70f..cd8b43097c 100644
--- a/tp/Texinfo/XS/convert/indices_in_conversion.c
+++ b/tp/Texinfo/XS/convert/indices_in_conversion.c
@@ -26,6 +26,7 @@
#include "utils.h"
#include "extra.h"
#include "unicode.h"
+#include "convert_to_text.h"
#include "indices_in_conversion.h"
/* corresponding perl code in Texinfo::Structuring */
@@ -160,3 +161,57 @@ index_content_element (const ELEMENT *element, int
prefer_reference_element)
}
}
+char *
+index_entry_element_sort_string (INDEX_ENTRY *main_entry,
+ ELEMENT *index_entry_element,
+ TEXT_OPTIONS *options,
+ int prefer_reference_element)
+{
+ char *sort_string;
+ char *index_ignore_chars;
+ ELEMENT *entry_tree_element;
+
+ if (!index_entry_element)
+ {
+ fatal ("index_entry_element_sort_string: NUL element");
+ }
+
+ char *sortas = lookup_extra_string (index_entry_element, "sortas");
+ if (sortas)
+ return strdup (sortas);
+
+ entry_tree_element = index_content_element (index_entry_element,
+ prefer_reference_element);
+
+ sort_string = convert_to_text (entry_tree_element, options);
+
+ index_ignore_chars = lookup_extra_string (main_entry->entry_element,
+ "index_ignore_chars");
+ if (index_ignore_chars)
+ {
+ TEXT sort_string_text;
+ char *p = sort_string;
+ text_init (&sort_string_text);
+
+ while (*p)
+ {
+ int n = strspn (p, index_ignore_chars);
+ if (n)
+ {
+ p += n;
+ }
+ if (*p)
+ {
+ /* store a character */
+ int char_len = 1;
+ while ((p[char_len] & 0xC0) == 0x80)
+ char_len++;
+ text_append_n (&sort_string_text, p, char_len);
+ p += char_len;
+ }
+ }
+ free (sort_string);
+ sort_string = sort_string_text.text;
+ }
+ return sort_string;
+}
diff --git a/tp/Texinfo/XS/convert/indices_in_conversion.h
b/tp/Texinfo/XS/convert/indices_in_conversion.h
index f516695632..7d9c7f7ab0 100644
--- a/tp/Texinfo/XS/convert/indices_in_conversion.h
+++ b/tp/Texinfo/XS/convert/indices_in_conversion.h
@@ -3,7 +3,7 @@
#define INDICES_IN_CONVERSION_H
#include "tree_types.h"
-#include "indices_in_conversion.h"
+#include "convert_to_text.h"
MERGED_INDEX *merge_indices (INDEX **index_names);
void destroy_merged_indices (MERGED_INDEX *merged_indices);
@@ -14,4 +14,8 @@ void destroy_indices_sorted_by_letter (
ELEMENT *index_content_element (const ELEMENT *element,
int prefer_reference_element);
+char *index_entry_element_sort_string (INDEX_ENTRY *main_entry,
+ ELEMENT *index_entry_element,
+ TEXT_OPTIONS *options,
+ int prefer_reference_element);
#endif
diff --git a/tp/Texinfo/XS/main/debug.c b/tp/Texinfo/XS/main/debug.c
index c925d44deb..43d72c9d65 100644
--- a/tp/Texinfo/XS/main/debug.c
+++ b/tp/Texinfo/XS/main/debug.c
@@ -99,7 +99,8 @@ print_element_debug (const ELEMENT *e, int print_parent)
return result;
}
-char *print_associate_info_debug (const ASSOCIATED_INFO *info)
+char *
+print_associate_info_debug (const ASSOCIATED_INFO *info)
{
TEXT text;
char *result;
diff --git a/tp/Texinfo/XS/main/get_perl_info.c
b/tp/Texinfo/XS/main/get_perl_info.c
index 6727421db1..984bca7ac0 100644
--- a/tp/Texinfo/XS/main/get_perl_info.c
+++ b/tp/Texinfo/XS/main/get_perl_info.c
@@ -41,6 +41,9 @@ FIXME add an initialization of translations?
#include "options_types.h"
#include "document_types.h"
#include "converter_types.h"
+#include "text.h"
+#include "extra.h"
+#include "debug.h"
#include "utils.h"
#include "builtin_commands.h"
#include "errors.h"
@@ -51,6 +54,63 @@ FIXME add an initialization of translations?
#include "converter.h"
#include "get_perl_info.h"
+#define FETCH(key) key##_sv = hv_fetch (element_hv, #key, strlen(#key), 0);
+
+static void
+debug_print_element_hv (HV *element_hv)
+{
+ SV **cmdname_sv;
+ SV **type_sv;
+ SV **text_sv;
+ TEXT msg;
+
+ dTHX;
+
+ text_init (&msg);
+ text_append (&msg, "");
+
+ FETCH(cmdname)
+ if (cmdname_sv)
+ {
+ text_printf (&msg, "@%s", SvPVutf8_nolen (*cmdname_sv));
+ }
+ FETCH(type)
+ if (type_sv)
+ {
+ text_printf (&msg, "(%s)", SvPVutf8_nolen (*type_sv));
+ }
+ FETCH(text)
+ if (text_sv)
+ {
+ int allocated = 0;
+ char *text = SvPVutf8_nolen (*text_sv);
+ char *protected_text = debug_protect_eol (text,
+ &allocated);
+ text_printf (&msg, "[T: %s]", protected_text);
+ if (allocated)
+ free (protected_text);
+ }
+ fprintf (stderr, "ELT_sv: %s\n", msg.text);
+ free (msg.text);
+}
+
+void
+debug_print_element_sv (SV *element_sv)
+{
+ dTHX;
+
+ if (SvOK (element_sv))
+ {
+ HV *element_hv = (HV *) SvRV (element_sv);
+ debug_print_element_hv (element_hv);
+ }
+ else
+ {
+ fprintf(stderr, "debug_print_element_sv: NUL\n");
+ }
+}
+#undef FETCH
+
DOCUMENT *
get_document_or_warn (SV *sv_in, char *key, char *warn_string)
{
@@ -336,7 +396,7 @@ get_expanded_formats (HV *hv, EXPANDED_FORMAT
**expanded_formats)
expanded_formats_sv = hv_fetch (hv, "expanded_formats",
strlen ("expanded_formats"), 0);
- if (expanded_formats_sv)
+ if (expanded_formats_sv && SvOK (*expanded_formats_sv))
{
I32 i;
I32 formats_nr;
@@ -573,6 +633,54 @@ reset_output_init_conf (SV *sv_in)
}
}
+INDEX_ENTRY *
+find_index_entry_sv (SV *index_entry_sv, INDEX **index_names,
+ const char *warn_string, char **entry_index_name,
+ int *entry_number)
+{
+ HV *index_entry_hv;
+ SV **index_name_sv;
+ SV **entry_number_sv;
+ int entry_idx_in_index;
+ INDEX *idx;
+
+ dTHX;
+
+ index_entry_hv = (HV *) SvRV (index_entry_sv);
+ index_name_sv = hv_fetch (index_entry_hv, "index_name",
+ strlen ("index_name"), 0);
+ entry_number_sv = hv_fetch (index_entry_hv, "entry_number",
+ strlen ("entry_number"), 0);
+
+ *entry_index_name = 0;
+ *entry_number = 0;
+
+ if (!index_name_sv || !entry_number_sv)
+ {
+ char *msg;
+ const char *warn_str = warn_string;
+ if (!warn_str)
+ warn_str = "find_index_entry_sv";
+ xasprintf (&msg, "%s: no entry info\n", warn_str);
+ fatal (msg);
+ }
+ *entry_index_name = (char *) SvPVutf8_nolen (*index_name_sv);
+ *entry_number = SvIV (*entry_number_sv);
+ entry_idx_in_index = *entry_number - 1;
+
+ idx = indices_info_index_by_name (index_names,
+ *entry_index_name);
+
+ if (idx)
+ {
+ if (entry_idx_in_index < idx->entries_number)
+ return &idx->index_entries[entry_idx_in_index];
+ }
+
+ return 0;
+}
+
+
/* code in comments allow to sort the index names to have a fixed order
in the data structure. Not clear that it is useful or not, not enabled
for now */
@@ -736,15 +844,10 @@ get_sv_index_entries_sorted_by_letter (INDEX
**index_names,
for (k = 0; k < entries_nr; k++)
{
SV** index_entry_sv = av_fetch (entries_av, k, 0);
- HV *index_entry_hv;
- SV** index_name_sv;
- SV** entry_number_sv;
- INDEX *idx;
char *entry_index_name;
int entry_number;
- int entry_idx_in_index;
-
- letter_entries->entries[k] = 0;
+ char *warn_string;
+ INDEX_ENTRY *index_entry = 0;
if (!index_entry_sv)
{
@@ -754,32 +857,16 @@ get_sv_index_entries_sorted_by_letter (INDEX
**index_names,
idx_name, i, letter_entries->letter, k);
fatal (msg);
}
- index_entry_hv = (HV *) SvRV (*index_entry_sv);
- index_name_sv = hv_fetch (index_entry_hv, "index_name",
- strlen ("index_name"), 0);
- entry_number_sv = hv_fetch (index_entry_hv, "entry_number",
- strlen ("entry_number"), 0);
- if (!index_name_sv || !entry_number_sv)
- {
- char *msg;
- xasprintf (&msg,
- "get_sv_index_entries_sorted_by_letter: %s: %d: %s: %d: no entry info\n",
- idx_name, i, letter_entries->letter, k);
- fatal (msg);
- }
- entry_index_name = (char *) SvPVutf8_nolen (*index_name_sv);
- entry_number = SvIV (*entry_number_sv);
- entry_idx_in_index = entry_number - 1;
+ xasprintf (&warn_string,
+ "get_sv_index_entries_sorted_by_letter: %s: %d: %s:
%d",
+ idx_name, i, letter_entries->letter, k);
+ index_entry = find_index_entry_sv (*index_entry_sv,
index_names,
+ warn_string,
&entry_index_name,
+ &entry_number);
+ free (warn_string);
- idx = indices_info_index_by_name (index_names,
- entry_index_name);
+ letter_entries->entries[k] = index_entry;
- if (idx)
- {
- if (entry_idx_in_index < idx->entries_number)
- letter_entries->entries[k]
- = &idx->index_entries[entry_idx_in_index];
- }
if (!letter_entries->entries[k])
{
char *msg;
@@ -824,12 +911,13 @@ force_conf (CONVERTER *converter, const char *conf, SV
*value)
/* output format specific */
/* map hash reference of Convert::Text options to TEXT_OPTIONS */
-/* TODO more to do */
+/* TODO more to do? */
#define FETCH(key) key##_sv = hv_fetch (hv_in, #key, strlen(#key), 0);
TEXT_OPTIONS *
copy_sv_options_for_convert_text (SV *sv_in)
{
HV *hv_in;
+ SV **code_sv;
SV **TEST_sv;
SV **INCLUDE_DIRECTORIES_sv;
SV **converter_sv;
@@ -853,8 +941,11 @@ copy_sv_options_for_convert_text (SV *sv_in)
if (enabled_encoding_sv)
text_options->encoding = strdup (SvPVutf8_nolen (*enabled_encoding_sv));
- FETCH(INCLUDE_DIRECTORIES)
+ FETCH(code)
+ if (code_sv)
+ text_options->code_state = SvIV (*code_sv);
+ FETCH(INCLUDE_DIRECTORIES)
if (INCLUDE_DIRECTORIES_sv)
add_svav_to_string_list (*INCLUDE_DIRECTORIES_sv,
&text_options->include_directories, svt_dir);
@@ -1145,6 +1236,7 @@ find_document_index_entry_extra_index_entry_sv (DOCUMENT
*document,
{
AV *extra_index_entry_av;
SV **index_name_sv;
+ char *index_name = 0;
INDEX *idx = 0;
dTHX;
@@ -1157,7 +1249,7 @@ find_document_index_entry_extra_index_entry_sv (DOCUMENT
*document,
index_name_sv = av_fetch (extra_index_entry_av, 0, 0);
if (index_name_sv)
{
- char *index_name = SvPVutf8_nolen (*index_name_sv);
+ index_name = SvPVutf8_nolen (*index_name_sv);
idx = indices_info_index_by_name (document->index_names,
index_name);
}
@@ -1178,7 +1270,7 @@ find_document_index_entry_extra_index_entry_sv (DOCUMENT
*document,
/* if there is a converter with sorted index names, use the
sorted index names, otherwise use the index information from
a document */
-ELEMENT *
+static INDEX_ENTRY *
find_element_extra_index_entry_sv (DOCUMENT *document,
CONVERTER *converter,
SV *extra_index_entry_sv)
@@ -1197,14 +1289,7 @@ find_element_extra_index_entry_sv (DOCUMENT *document,
index_entry = find_sorted_index_names_index_entry_extra_index_entry_sv (
&converter->sorted_index_names, extra_index_entry_sv);
- if (index_entry)
- {
- if (index_entry->entry_associated_element)
- return index_entry->entry_associated_element;
- else if (index_entry->entry_element)
- return index_entry->entry_element;
- }
- return 0;
+ return index_entry;
}
#define FETCH(key) key##_sv = hv_fetch (element_hv, #key, strlen(#key), 0);
@@ -1262,6 +1347,111 @@ ELEMENT *find_root_command (DOCUMENT *document, HV
*element_hv,
return 0;
}
+/* find the subentry matching ELEMENT_HV */
+static ELEMENT *
+find_index_entry_subentry (ELEMENT *index_element, HV *element_hv)
+{
+ ELEMENT *current_element = index_element;
+
+ while (1)
+ {
+ ELEMENT *subentry = lookup_extra_element (current_element,
+ "subentry");
+ if (subentry)
+ {
+ if (subentry->hv == element_hv)
+ return subentry;
+ current_element = subentry;
+ }
+ else
+ return 0;
+ }
+}
+
+#define EXTRA(key) key##_sv = hv_fetch (extra_hv, #key, strlen(#key), 0);
+
+/* returns the subentry direct parent based on "subentry_parent" */
+static SV *
+subentry_hv_parent (HV *element_hv)
+{
+ SV **extra_sv;
+
+ dTHX;
+
+ FETCH(extra)
+
+ if (extra_sv)
+ {
+ SV **subentry_parent_sv;
+ HV *extra_hv = (HV *) SvRV (*extra_sv);
+
+ EXTRA(subentry_parent)
+ if (subentry_parent_sv)
+ {
+ return *subentry_parent_sv;
+ }
+ }
+ return 0;
+}
+
+/* Find the index entry parent of a subentry going through
+ "subentry_parent" until finding the index element hash */
+ELEMENT *
+find_subentry_index_command_sv (DOCUMENT *document, HV *element_hv)
+{
+ HV *current_parent = element_hv;
+ SV *current_sv = 0;
+
+ dTHX;
+
+ while (1)
+ {
+ SV *subentry_parent_sv = subentry_hv_parent (current_parent);
+ if (subentry_parent_sv)
+ {
+ current_parent = (HV *) SvRV (subentry_parent_sv);
+ current_sv = subentry_parent_sv;
+ }
+ else
+ {
+ if (!current_sv)
+ return 0;
+ return find_element_from_sv (0, document, current_sv, 0);
+ }
+ }
+}
+
+/* find the INDEX_ENTRY associated element matching ELEMENT_HV.
+
+ If the index entry was reassociated, the tree element the
+ index entry is reassociated to is not index_entry->entry_element
+ but index_entry->entry_associated_element. The original
+ tree element that was associated is index_entry->entry_element.
+ Depending on the situation one or the other may be looked for
+ and the code tries both.
+
+ The reassociated tree element, for example, would be used
+ when doing a link to the tree from the index entry. But it may
+ also be the original tree element that is used, for example
+ to get the index entry tree element content, for instance
+ when going through the elements associated to indices to setup
+ index entries sort strings.
+ */
+ELEMENT *find_index_entry_associated_hv (INDEX_ENTRY *index_entry,
+ HV *element_hv)
+{
+ if (index_entry->entry_associated_element
+ && index_entry->entry_associated_element->hv == element_hv)
+ return index_entry->entry_associated_element;
+
+ if (index_entry->entry_element
+ /* if the index entry was reassociated it is important to check */
+ && index_entry->entry_element->hv == element_hv)
+ return index_entry->entry_element;
+
+ return 0;
+}
+
/* TODO nodedescription using the extra element_node and the
* node extra node_description? */
@@ -1289,11 +1479,11 @@ find_element_from_sv (CONVERTER *converter, DOCUMENT
*document_in,
element_hv = (HV *) SvRV (element_sv);
- FETCH(cmdname)
-
if (!document && converter && converter->document)
document = converter->document;
+ FETCH(cmdname)
+
if (cmdname_sv && (output_units_descriptor || document))
{
char *cmdname = SvPVutf8_nolen (*cmdname_sv);
@@ -1308,11 +1498,22 @@ find_element_from_sv (CONVERTER *converter, DOCUMENT
*document_in,
if (element)
return element;
}
+ else if (cmd == CM_subentry)
+ {
+ ELEMENT *index_element = find_subentry_index_command_sv (document,
+ element_hv);
+ if (index_element)
+ {
+ ELEMENT *element = find_index_entry_subentry (index_element,
+ element_hv);
+ if (element)
+ return element;
+ }
+ }
}
FETCH(extra)
-#define EXTRA(key) key##_sv = hv_fetch (extra_hv, #key, strlen(#key), 0);
if (extra_sv)
{
HV *extra_hv = (HV *) SvRV (*extra_sv);
@@ -1352,27 +1553,36 @@ find_element_from_sv (CONVERTER *converter, DOCUMENT
*document_in,
}
}
-
EXTRA(associated_index_entry)
if (associated_index_entry_sv)
{
- ELEMENT *index_element = find_element_extra_index_entry_sv (document,
- converter,
- *associated_index_entry_sv);
- /* there should be no ambiguity, but we check nevertheless */
- if (index_element && index_element->hv == element_hv)
- return (index_element);
+ INDEX_ENTRY *index_entry
+ = find_element_extra_index_entry_sv (document,
+ converter,
+ *associated_index_entry_sv);
+ if (index_entry)
+ {
+ ELEMENT *index_element
+ = find_index_entry_associated_hv (index_entry, element_hv);
+ if (index_element)
+ return (index_element);
+ }
}
EXTRA(index_entry)
if (index_entry_sv)
{
- ELEMENT *index_element = find_element_extra_index_entry_sv (document,
+ INDEX_ENTRY *index_entry
+ = find_element_extra_index_entry_sv (document,
converter,
*index_entry_sv);
- /* it is important to check if the index entry was reassociated */
- if (index_element && index_element->hv == element_hv)
- return (index_element);
+ if (index_entry)
+ {
+ ELEMENT *index_element
+ = find_index_entry_associated_hv (index_entry, element_hv);
+ if (index_element)
+ return (index_element);
+ }
}
}
return 0;
diff --git a/tp/Texinfo/XS/main/get_perl_info.h
b/tp/Texinfo/XS/main/get_perl_info.h
index 8276e43f27..5be575eff5 100644
--- a/tp/Texinfo/XS/main/get_perl_info.h
+++ b/tp/Texinfo/XS/main/get_perl_info.h
@@ -42,6 +42,10 @@ CONVERTER *get_sv_converter (SV *sv_in, const char
*warn_string);
int converter_initialize (SV *converter_sv);
void reset_output_init_conf (SV *sv_in);
+INDEX_ENTRY *find_index_entry_sv (SV *index_entry_sv, INDEX **index_names,
+ const char *warn_string, char **entry_index_name,
+ int *entry_number);
+
INDEX_SORTED_BY_LETTER *get_sv_index_entries_sorted_by_letter
(INDEX **index_names, SV *index_entries_sorted_by_letter);
diff --git a/tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
b/tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
index d1cad4a903..900f9e9c22 100644
--- a/tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
+++ b/tp/Texinfo/XS/structuring_transfo/StructuringTransfoXS.xs
@@ -40,6 +40,7 @@
#include "transformations.h"
#include "structuring.h"
#include "output_unit.h"
+#include "indices_in_conversion.h"
#include "get_perl_info.h"
#include "build_perl_info.h"
@@ -437,4 +438,37 @@ split_pages (SV *output_units_in, char *split)
if (output_units)
split_pages (output_units, split);
+SV *
+index_entry_element_sort_string (SV *document_in, SV *main_entry_sv, SV
*element_sv, SV *options_sv, SV *prefer_reference_element_sv=0)
+ PREINIT:
+ DOCUMENT *document;
+ char *sort_string = 0;
+ CODE:
+ document = get_sv_document_document (document_in,
+ "index_entry_element_sort_string");
+ if (document)
+ {
+ char *entry_index_name;
+ int entry_number;
+ int prefer_reference_element = 0;
+ ELEMENT *element = find_element_from_sv (0, document,
+ element_sv, 0);
+ INDEX_ENTRY *main_entry = find_index_entry_sv (main_entry_sv,
+ document->index_names, 0,
+ &entry_index_name, &entry_number);
+ TEXT_OPTIONS *options
+ = copy_sv_options_for_convert_text (options_sv);
+ if (prefer_reference_element_sv && SvOK
(prefer_reference_element_sv))
+ prefer_reference_element = SvIV (prefer_reference_element_sv);
+ sort_string = index_entry_element_sort_string (main_entry,
+ element, options, prefer_reference_element);
+ destroy_text_options (options);
+ }
+
+ if (!sort_string)
+ RETVAL = newSV (0);
+ else
+ RETVAL = newSVpv_utf8 (sort_string, 0);
+ OUTPUT:
+ RETVAL
diff --git a/tp/t/test_sort.t b/tp/t/test_sort.t
index fac7006334..eef08d15c0 100644
--- a/tp/t/test_sort.t
+++ b/tp/t/test_sort.t
@@ -49,6 +49,8 @@ my $index_entries =
Texinfo::Structuring::merge_indices($indices_information);
my $document_information = $document->global_information();
my $main_configuration = Texinfo::MainConfig::new({'ENABLE_ENCODING' => 1});
Texinfo::Common::set_output_encodings($main_configuration,
$document_information);
+$main_configuration->{'document_descriptor'}
+ = $document->document_descriptor();
my ($sorted_index_entries, $index_entries_sort_strings)
= Texinfo::Structuring::sort_indices_by_index($registrar,
$main_configuration,
$index_entries,
$indices_information);
@@ -124,6 +126,8 @@ $tree = $document->tree();
$registrar = $parser->registered_errors();
$indices_information = $document->indices_information();
$index_entries = Texinfo::Structuring::merge_indices($indices_information);
+$main_configuration->{'document_descriptor'}
+ = $document->document_descriptor();
($sorted_index_entries, $index_entries_sort_strings)
= Texinfo::Structuring::sort_indices_by_index($registrar,
$main_configuration,
$index_entries,
$indices_information);
diff --git a/tp/t/test_utils.pl b/tp/t/test_utils.pl
index 00930006bc..560dab6dd3 100644
--- a/tp/t/test_utils.pl
+++ b/tp/t/test_utils.pl
@@ -1177,6 +1177,8 @@ sub test($$)
my ($sorted_index_entries, $index_entries_sort_strings);
my $indices_sorted_sort_strings;
if ($merged_index_entries) {
+ $main_configuration->{'document_descriptor'}
+ = $document->document_descriptor();
($sorted_index_entries, $index_entries_sort_strings)
= Texinfo::Structuring::sort_indices_by_index($registrar,
$main_configuration,
| [Prev in Thread] |
Current Thread |
[Next in Thread] |
- branch master updated: * tp/Texinfo/Structuring.pm (setup_index_entry_keys_formatting): remove ascii_punctuation obsolete option set.,
Patrice Dumas <=