lib/libxml2
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/libxml2.git synced 2025-07-29 11:41:22 +03:00
Code Activity

Better entities and char ref encoding, error msg formatting, Daniel.

Browse Source
This commit is contained in:
Daniel Veillard
1998-11-16 01:04:26 +00:00
parent 242590ee5d
commit 8cc0d1f889
3 changed files with 49 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
entities.c
View File

@ -248,6 +248,15 @@ xmlEntityPtr xmlGetDocEntity(xmlDocPtr doc, const CHAR *name) {
return(NULL);
}
/*
* [2] Char ::= #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD]
* | [#x10000-#x10FFFF]
* any Unicode character, excluding the surrogate blocks, FFFE, and FFFF.
*/
#define IS_CHAR(c) \
(((c) == 0x09) || ((c) == 0x0a) || ((c) == 0x0d) || \
(((c) >= 0x20) && ((c) != 0xFFFE) && ((c) != 0xFFFF)))
/*
* xmlEncodeEntities : do a global encoding of a string, replacing the
* predefined entities and non ASCII values with their
@ -255,6 +264,7 @@ xmlEntityPtr xmlGetDocEntity(xmlDocPtr doc, const CHAR *name) {
* TODO !!!! Once moved to UTF-8 internal encoding, the encoding of non-ascii
* get erroneous.
*/
CHAR *xmlEncodeEntities(xmlDocPtr doc, const CHAR *input) {
const CHAR *cur = input;
CHAR *out = buffer;
@ -310,6 +320,12 @@ CHAR *xmlEncodeEntities(xmlDocPtr doc, const CHAR *input) {
*out++ = 'o';
*out++ = 's';
*out++ = ';';
} else if (((*cur >= 0x20) && (*cur < 0x80)) ||
(*cur == '\n') || (*cur == '\r') || (*cur == '\t')) {
/*
* default case, just copy !
*/
*out++ = *cur;
#ifndef USE_UTF_8
} else if ((sizeof(CHAR) == 1) && (*cur >= 0x80)) {
char buf[10], *ptr;
@ -321,12 +337,26 @@ CHAR *xmlEncodeEntities(xmlDocPtr doc, const CHAR *input) {
ptr = buf;
while (*ptr != 0) *out++ = *ptr++;
#endif
} else {
/*
* default case, just copy !
*/
*out++ = *cur;
} else if (IS_CHAR(*cur)) {
char buf[10], *ptr;
#ifdef HAVE_SNPRINTF
snprintf(buf, 9, "&#%d;", *cur);
#else
sprintf(buf, "&#%d;", *cur);
#endif
ptr = buf;
while (*ptr != 0) *out++ = *ptr++;
}
#if 0
else {
/*
* default case, this is not a valid char !
* Skip it...
*/
fprintf(stderr, "xmlEncodeEntities: invalid char %d\n", (int) *cur);
}
#endif
cur++;
}
*out++ = 0;