[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[no subject]
|
From: |
Patrice Dumas |
|
Date: |
Wed, 15 Nov 2023 03:37:20 -0500 (EST) |
branch: master
commit 2f3d653e295562dc75de9654bee0feda2e497308
Author: Patrice Dumas <pertusus@free.fr>
AuthorDate: Wed Nov 15 09:34:18 2023 +0100
* tp/Texinfo/XS/convert/converter.h (OTXI_PROTECT_XML_CASES),
tp/Texinfo/XS/convert/converter.c
(xml_format_text_with_numeric_entities, xml_protect_text):
add code that can be inlined for the protection of XML special
characters as OTXI_PROTECT_XML_CASES. Use in converter.c and base
code more on misc.c code with memcmp and case.
* tp/Texinfo/XS/convert/convert_html.c
(OTXI_PROTECT_XML_FORM_FEED_CASES, html_default_format_protect_text)
(protect_text_use_iso_entities, protect_text_no_iso_entities):
setup OTXI_PROTECT_XML_FORM_FEED_CASES to be inlined for the
protection of XML special charcters and form feeds. Implement
protect_text_use_iso_entities and html_default_format_protect_text.
---
ChangeLog | 30 +++++--
tp/Texinfo/XS/convert/convert_html.c | 149 +++++++++++++++++++++++++++++++----
tp/Texinfo/XS/convert/converter.c | 93 +++++++++++-----------
tp/Texinfo/XS/convert/converter.h | 14 ++++
4 files changed, 218 insertions(+), 68 deletions(-)
diff --git a/ChangeLog b/ChangeLog
index 0a7210931f..059d8491e7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,12 +1,18 @@
-2023-11-14 Gavin Smith <gavinsmith0123@gmail.com>
+2023-11-15 Patrice Dumas <pertusus@free.fr>
- * tp/Texinfo/command_data.txt
- (item_LINE, itemx, defblock, defline, deftypeline):
- Remove contain_basic_inline flag. There is no reason an @anchor
- should not occur inside @item, inside @table, or the other
- commands, as no index entry is being created with the @anchor.
+ * tp/Texinfo/XS/convert/converter.h (OTXI_PROTECT_XML_CASES),
+ tp/Texinfo/XS/convert/converter.c
+ (xml_format_text_with_numeric_entities, xml_protect_text):
+ add code that can be inlined for the protection of XML special
+ characters as OTXI_PROTECT_XML_CASES. Use in converter.c and base
+ code more on misc.c code with memcmp and case.
- Report from Ihor Radchenko <yantar92@posteo.net> for Org mode manual.
+ * tp/Texinfo/XS/convert/convert_html.c
+ (OTXI_PROTECT_XML_FORM_FEED_CASES, html_default_format_protect_text)
+ (protect_text_use_iso_entities, protect_text_no_iso_entities):
+ setup OTXI_PROTECT_XML_FORM_FEED_CASES to be inlined for the
+ protection of XML special charcters and form feeds. Implement
+ protect_text_use_iso_entities and html_default_format_protect_text.
2023-11-14 Patrice Dumas <pertusus@free.fr>
@@ -31,6 +37,16 @@
* tp/Texinfo/XS/convert/converter.c
(xml_format_text_with_numeric_entities, xml_protect_text): implement.
+2023-11-14 Gavin Smith <gavinsmith0123@gmail.com>
+
+ * tp/Texinfo/command_data.txt
+ (item_LINE, itemx, defblock, defline, deftypeline):
+ Remove contain_basic_inline flag. There is no reason an @anchor
+ should not occur inside @item, inside @table, or the other
+ commands, as no index entry is being created with the @anchor.
+
+ Report from Ihor Radchenko <yantar92@posteo.net> for Org mode manual.
+
2023-11-14 Patrice Dumas <pertusus@free.fr>
* tp/Texinfo/Translations.pm (import), tp/Texinfo/XS/Makefile.am,
diff --git a/tp/Texinfo/XS/convert/convert_html.c
b/tp/Texinfo/XS/convert/convert_html.c
index 3f750f7881..958ee64405 100644
--- a/tp/Texinfo/XS/convert/convert_html.c
+++ b/tp/Texinfo/XS/convert/convert_html.c
@@ -2220,7 +2220,12 @@ html_prepare_units_directions_files (CONVERTER *self,
}
-#define ADDN(str,nr) text_append_n (result, str, nr)
+#define OTXI_PROTECT_XML_FORM_FEED_CASES(var) \
+ OTXI_PROTECT_XML_CASES(var) \
+ case '\f': \
+ text_append_n (result, "", 5); var++; \
+ break;
+
void
html_default_format_protect_text (const char *text, TEXT *result)
{
@@ -2238,26 +2243,13 @@ html_default_format_protect_text (const char *text,
TEXT *result)
break;
switch (*p)
{
- case '<':
- ADDN("<", 4);
- break;
- case '>':
- ADDN(">", 4);
- break;
- case '&':
- ADDN("&", 5);
- break;
- case '"':
- ADDN(""", 6);
- break;
- case '\f':
- ADDN("", 5);
- break;
+ OTXI_PROTECT_XML_FORM_FEED_CASES(p)
}
- p++;
}
}
+#define ADDN(str,nr) text_append_n (result, str, nr)
+
void
default_css_string_format_protect_text (const char *text, TEXT *result)
{
@@ -2286,6 +2278,129 @@ default_css_string_format_protect_text (const char
*text, TEXT *result)
}
}
+void
+protect_text_use_iso_entities (const char *text, TEXT *result)
+{
+ const char *p = text;
+
+ while (*p)
+ {
+ int before_sep_nr = strcspn (p, "<>&\"\f" "-`'");
+ if (before_sep_nr)
+ {
+ text_append_n (result, p, before_sep_nr);
+ p += before_sep_nr;
+ }
+ if (!*p)
+ break;
+ switch (*p)
+ {
+ OTXI_PROTECT_XML_FORM_FEED_CASES(p)
+ case '-':
+ if (*(p+1) && !memcmp (p, "---", 3))
+ {
+ ADDN("—", 7);
+ p += 3;
+ }
+ else if (!memcmp (p, "--", 2))
+ {
+ ADDN("–", 7);
+ p += 2;
+ }
+ else
+ {
+ ADDN("-", 1);
+ p++;
+ }
+ break;
+ case '`':
+ if (!memcmp (p, "``", 2))
+ {
+ ADDN("“", 7);
+ p += 2;
+ }
+ else
+ {
+ ADDN("‘", 7);
+ p++;
+ }
+ break;
+ case '\'':
+ if (!memcmp (p, "''", 2))
+ {
+ ADDN("”", 7);
+ p += 2;
+ }
+ else
+ {
+ ADDN("’", 7);
+ p++;
+ }
+ break;
+ }
+ }
+}
+
+void
+protect_text_no_iso_entities (const char *text, TEXT *result)
+{
+ const char *p = text;
+
+ while (*p)
+ {
+ int before_sep_nr = strcspn (p, "<>&\"\f" "-`'");
+ if (before_sep_nr)
+ {
+ text_append_n (result, p, before_sep_nr);
+ p += before_sep_nr;
+ }
+ if (!*p)
+ break;
+ switch (*p)
+ {
+ OTXI_PROTECT_XML_FORM_FEED_CASES(p)
+ case '-':
+ if (*(p+1) && !memcmp (p, "---", 3))
+ {
+ ADDN("--", 2);
+ p += 3;
+ }
+ else
+ {
+ if (!memcmp (p, "--", 2))
+ p += 2;
+ else
+ p++;
+ ADDN("-", 1);
+ }
+ break;
+ case '`':
+ if (!memcmp (p, "``", 2))
+ {
+ ADDN(""", 6);
+ p += 2;
+ }
+ else
+ {
+ ADDN(p, 1);
+ p++;
+ }
+ break;
+ case '\'':
+ if (!memcmp (p, "''", 2))
+ {
+ ADDN(""", 6);
+ p += 2;
+ }
+ else
+ {
+ ADDN(p, 1);
+ p++;
+ }
+ break;
+ }
+ }
+}
#undef ADDN
static char *
diff --git a/tp/Texinfo/XS/convert/converter.c
b/tp/Texinfo/XS/convert/converter.c
index dfe58b34f8..13d9246935 100644
--- a/tp/Texinfo/XS/convert/converter.c
+++ b/tp/Texinfo/XS/convert/converter.c
@@ -589,7 +589,6 @@ void
xml_format_text_with_numeric_entities (const char *text, TEXT *result)
{
const char *p;
- int str_len;
p = text;
while (*p)
@@ -602,40 +601,56 @@ xml_format_text_with_numeric_entities (const char *text,
TEXT *result)
}
if (!*p)
break;
- str_len = strlen (p);
- if ((str_len > 1) && (!strncmp (p, "``", 2)))
- {
- ADD(8220);
- p += 2;
- }
- else if ((str_len > 1) && (!strncmp (p, "''", 2)))
- {
- ADD(8221);
- p += 2;
- }
- else if ((str_len > 2) && !strncmp (p, "---", 3))
- {
- ADD(8212);
- p += 3;
- }
- else if ((str_len > 1) && !strncmp (p, "--", 2))
- {
- ADD(8211);
- p += 2;
- }
- else
+ switch (*p)
{
- if (*p == '\'')
- ADD(8217);
- else if (*p == '`')
- ADD(8216);
- p++;
+ case '-':
+ if (*(p+1) && !memcmp (p, "---", 3))
+ {
+ ADD(8212);
+ p += 3;
+ }
+ else if (!memcmp (p, "--", 2))
+ {
+ ADD(8211);
+ p += 2;
+ }
+ else
+ {
+ text_append_n (result, "-", 1);
+ p++;
+ }
+ break;
+ case '`':
+ if (!memcmp (p, "``", 2))
+ {
+ ADD(8220);
+ p += 2;
+ }
+ else
+ {
+ ADD(8216);
+ p++;
+ }
+ break;
+ case '\'':
+ if (!memcmp (p, "''", 2))
+ {
+ ADD(8221);
+ p += 2;
+ }
+ else
+ {
+ ADD(8217);
+ p++;
+ }
+ break;
}
}
}
#undef ADD
-#define ADDN(str,nr) text_append_n (result, str, nr)
+
+
void
xml_protect_text (const char *text, TEXT *result)
{
@@ -645,7 +660,7 @@ xml_protect_text (const char *text, TEXT *result)
while (*p)
{
- int before_sep_nr = strcspn (p, "<>&\"\f");
+ int before_sep_nr = strcspn (p, "<>&\"");
if (before_sep_nr)
{
text_append_n (result, p, before_sep_nr);
@@ -655,20 +670,10 @@ xml_protect_text (const char *text, TEXT *result)
break;
switch (*p)
{
- case '<':
- ADDN("<", 4);
- break;
- case '>':
- ADDN(">", 4);
- break;
- case '&':
- ADDN("&", 5);
- break;
- case '"':
- ADDN(""", 6);
- break;
+ OTXI_PROTECT_XML_CASES(p);
+ /* should never happen */
+ default:
+ p++;
}
- p++;
}
}
-#undef ADDN
diff --git a/tp/Texinfo/XS/convert/converter.h
b/tp/Texinfo/XS/convert/converter.h
index c864784f47..fd878f6fa5 100644
--- a/tp/Texinfo/XS/convert/converter.h
+++ b/tp/Texinfo/XS/convert/converter.h
@@ -10,6 +10,20 @@
/* for TARGET_FILENAME */
#include "utils.h"
+#define OTXI_PROTECT_XML_CASES(var) \
+ case '<': \
+ text_append_n(result, "<", 4); var++; \
+ break; \
+ case '>': \
+ text_append_n (result, ">", 4); var++; \
+ break; \
+ case '&': \
+ text_append_n (result, "&", 5); var++; \
+ break; \
+ case '"': \
+ text_append_n (result, """, 6); var++; \
+ break;
+
CONVERTER *retrieve_converter (int converter_descriptor);
size_t register_converter (CONVERTER *converter);