mirror of
https://gitlab.gnome.org/GNOME/libxml2
synced 2025-03-28 21:33:13 +00:00
parser: Implement xmlCtxtSetOptions
Surprisingly, some options can only be enabled with xmlCtxtUseOptions and it's impossible to unset them. Add a new API function xmlCtxtSetOptions which sets or clears all options. Finally document all parser options. Make sure to synchronize option bits and struct members.
This commit is contained in:
Nick Wellnhofer
parent
33ec407a73
commit
875bb08489
@ -1272,6 +1272,9 @@ XMLPUBFUN int
|
||||
int size,
|
||||
const char *filename,
|
||||
const char *encoding);
|
||||
XMLPUBFUN int
|
||||
xmlCtxtSetOptions (xmlParserCtxtPtr ctxt,
|
||||
int options);
|
||||
XMLPUBFUN int
|
||||
xmlCtxtUseOptions (xmlParserCtxtPtr ctxt,
|
||||
int options);
|
||||
|
||||
374
parser.c
374
parser.c
@ -13341,13 +13341,257 @@ xmlCtxtResetPush(xmlParserCtxtPtr ctxt, const char *chunk,
|
||||
return(0);
|
||||
}
|
||||
|
||||
static int
|
||||
xmlCtxtSetOptionsInternal(xmlParserCtxtPtr ctxt, int options, int keepMask)
|
||||
{
|
||||
int allMask;
|
||||
|
||||
if (ctxt == NULL)
|
||||
return(-1);
|
||||
|
||||
/*
|
||||
* XInclude options aren't handled by the parser.
|
||||
*
|
||||
* XML_PARSE_XINCLUDE
|
||||
* XML_PARSE_NOXINCNODE
|
||||
* XML_PARSE_NOBASEFIX
|
||||
*/
|
||||
allMask = XML_PARSE_RECOVER |
|
||||
XML_PARSE_NOENT |
|
||||
XML_PARSE_DTDLOAD |
|
||||
XML_PARSE_DTDATTR |
|
||||
XML_PARSE_DTDVALID |
|
||||
XML_PARSE_NOERROR |
|
||||
XML_PARSE_NOWARNING |
|
||||
XML_PARSE_PEDANTIC |
|
||||
XML_PARSE_NOBLANKS |
|
||||
#ifdef LIBXML_SAX1_ENABLED
|
||||
XML_PARSE_SAX1 |
|
||||
#endif
|
||||
XML_PARSE_NONET |
|
||||
XML_PARSE_NODICT |
|
||||
XML_PARSE_NSCLEAN |
|
||||
XML_PARSE_NOCDATA |
|
||||
XML_PARSE_COMPACT |
|
||||
XML_PARSE_OLD10 |
|
||||
XML_PARSE_HUGE |
|
||||
XML_PARSE_OLDSAX |
|
||||
XML_PARSE_IGNORE_ENC |
|
||||
XML_PARSE_BIG_LINES;
|
||||
|
||||
ctxt->options = (ctxt->options & keepMask) | (options & allMask);
|
||||
|
||||
/*
|
||||
* For some options, struct members are historically the source
|
||||
* of truth. The values are initalized from global variables and
|
||||
* old code could also modify them directly. Several older API
|
||||
* functions that don't take an options argument rely on these
|
||||
* deprecated mechanisms.
|
||||
*
|
||||
* Once public access to struct members and the globals are
|
||||
* disabled, we can use the options bitmask as source of
|
||||
* truth, making all these struct members obsolete.
|
||||
*/
|
||||
ctxt->recovery = (options & XML_PARSE_RECOVER) ? 1 : 0;
|
||||
ctxt->replaceEntities = (options & XML_PARSE_NOENT) ? 1 : 0;
|
||||
ctxt->loadsubset = (options & XML_PARSE_DTDLOAD) ? XML_DETECT_IDS : 0;
|
||||
ctxt->loadsubset |= (options & XML_PARSE_DTDATTR) ? XML_COMPLETE_ATTRS : 0;
|
||||
ctxt->validate = (options & XML_PARSE_DTDVALID) ? 1 : 0;
|
||||
ctxt->pedantic = (options & XML_PARSE_PEDANTIC) ? 1 : 0;
|
||||
ctxt->keepBlanks = (options & XML_PARSE_NOBLANKS) ? 0 : 1;
|
||||
ctxt->dictNames = (options & XML_PARSE_NODICT) ? 0 : 1;
|
||||
|
||||
/*
|
||||
* Changing SAX callbacks is a bad idea. This should be fixed.
|
||||
*/
|
||||
if (options & XML_PARSE_NOBLANKS) {
|
||||
ctxt->sax->ignorableWhitespace = xmlSAX2IgnorableWhitespace;
|
||||
}
|
||||
if (options & XML_PARSE_NOCDATA) {
|
||||
ctxt->sax->cdataBlock = NULL;
|
||||
}
|
||||
if (options & XML_PARSE_HUGE) {
|
||||
if (ctxt->dict != NULL)
|
||||
xmlDictSetLimit(ctxt->dict, 0);
|
||||
}
|
||||
|
||||
ctxt->linenumbers = 1;
|
||||
|
||||
return(options & ~allMask);
|
||||
}
|
||||
|
||||
/**
|
||||
* xmlCtxtSetOptions:
|
||||
* @ctxt: an XML parser context
|
||||
* @options: a bitmask of xmlParserOption values
|
||||
*
|
||||
* Applies the options to the parser context. Unset options are
|
||||
* cleared.
|
||||
*
|
||||
* Available since 2.13.0. With older versions, you can use
|
||||
* xmlCtxtUseOptions.
|
||||
*
|
||||
* XML_PARSE_RECOVER
|
||||
*
|
||||
* Enable "recovery" mode which allows non-wellformed documents.
|
||||
* How this mode behaves exactly is unspecified and may change
|
||||
* without further notice. Use of this feature is DISCOURAGED.
|
||||
*
|
||||
* XML_PARSE_NOENT
|
||||
*
|
||||
* Despite the confusing name, this option enables substitution
|
||||
* of entities. The resulting tree won't contain any entity
|
||||
* reference nodes. This option also enables loading of
|
||||
* external entities which is dangerous. If you process
|
||||
* untrusted data, it's recommended to set up an external entity
|
||||
* loader that validates the files or URIs being loaded.
|
||||
*
|
||||
* This option also enables the loading and substitution of
|
||||
* external parameter entities. Internal parameter entities are
|
||||
* always expanded.
|
||||
*
|
||||
* XML_PARSE_DTDLOAD
|
||||
*
|
||||
* Enables loading of an external DTD and the loading and
|
||||
* substitution of external parameter entities.
|
||||
*
|
||||
* XML_PARSE_DTDATTR
|
||||
*
|
||||
* Adds default attributes from the DTD to the result document.
|
||||
*
|
||||
* Implies XML_PARSE_DTDLOAD.
|
||||
*
|
||||
* XML_PARSE_DTDVALID
|
||||
*
|
||||
* This option enables DTD validation.
|
||||
*
|
||||
* Implies XML_PARSE_DTDLOAD.
|
||||
*
|
||||
* XML_PARSE_NOERROR
|
||||
*
|
||||
* Disable error and warning reports to the error handlers.
|
||||
* Errors are still accessible with xmlCtxtGetLastError.
|
||||
*
|
||||
* XML_PARSE_NOWARNING
|
||||
*
|
||||
* Disable warning reports.
|
||||
*
|
||||
* XML_PARSE_PEDANTIC
|
||||
*
|
||||
* Enable some pedantic warnings.
|
||||
*
|
||||
* XML_PARSE_NOBLANKS
|
||||
*
|
||||
* Remove some text nodes containing only whitespace from the
|
||||
* result document. Which nodes are removed depends on DTD
|
||||
* element declarations or a conservative heuristic. The
|
||||
* reindenting feature of the serialization code relies on this
|
||||
* option to be set when parsing. Use of this option is
|
||||
* DISCOURAGED.
|
||||
*
|
||||
* XML_PARSE_SAX1
|
||||
*
|
||||
* Always invoke the deprecated SAX1 startElement and endElement
|
||||
* handlers. This option is DEPRECATED.
|
||||
*
|
||||
* XML_PARSE_NONET
|
||||
*
|
||||
* Disable network access with the builtin HTTP and FTP clients.
|
||||
*
|
||||
* XML_PARSE_NODICT
|
||||
*
|
||||
* Create a document without interned strings, making all
|
||||
* strings separate memory allocations.
|
||||
*
|
||||
* XML_PARSE_NSCLEAN
|
||||
*
|
||||
* Remove redundant namespace declarations from the result
|
||||
* document.
|
||||
*
|
||||
* XML_PARSE_NOCDATA
|
||||
*
|
||||
* Output normal text nodes instead of CDATA nodes.
|
||||
*
|
||||
* XML_PARSE_COMPACT
|
||||
*
|
||||
* Store small strings directly in the node struct to save
|
||||
* memory.
|
||||
*
|
||||
* XML_PARSE_OLD10
|
||||
*
|
||||
* Use old Name productions from before XML 1.0 Fifth Edition.
|
||||
* This options is DEPRECATED.
|
||||
*
|
||||
* XML_PARSE_HUGE
|
||||
*
|
||||
* Relax some internal limits.
|
||||
*
|
||||
* Maximum size of text nodes, tags, comments, processing instructions,
|
||||
* CDATA sections, entity values
|
||||
*
|
||||
* normal: 10M
|
||||
* huge: 1B
|
||||
*
|
||||
* Maximum size of names, system literals, pubid literals
|
||||
*
|
||||
* normal: 50K
|
||||
* huge: 10M
|
||||
*
|
||||
* Maximum nesting depth of elements
|
||||
*
|
||||
* normal: 256
|
||||
* huge: 2048
|
||||
*
|
||||
* Maximum nesting depth of entities
|
||||
*
|
||||
* normal: 20
|
||||
* huge: 40
|
||||
*
|
||||
* XML_PARSE_OLDSAX
|
||||
*
|
||||
* Enable an unspecified legacy mode for SAX parsers. This
|
||||
* option is DEPRECATED.
|
||||
*
|
||||
* XML_PARSE_IGNORE_ENC
|
||||
*
|
||||
* Ignore the encoding in the XML declaration. This option is
|
||||
* mostly unneeded these days. The only effect is to enforce
|
||||
* UTF-8 decoding of ASCII-like data.
|
||||
*
|
||||
* XML_PARSE_BIG_LINES
|
||||
*
|
||||
* Enable reporting of line numbers larger than 65535.
|
||||
*
|
||||
* Returns 0 in case of success, the set of unknown or unimplemented options
|
||||
* in case of error.
|
||||
*/
|
||||
int
|
||||
xmlCtxtSetOptions(xmlParserCtxtPtr ctxt, int options)
|
||||
{
|
||||
return(xmlCtxtSetOptionsInternal(ctxt, options, 0));
|
||||
}
|
||||
|
||||
/**
|
||||
* xmlCtxtUseOptions:
|
||||
* @ctxt: an XML parser context
|
||||
* @options: a combination of xmlParserOption
|
||||
*
|
||||
* Applies the options to the parser context
|
||||
* DEPRECATED: Use xmlCtxtSetOptions.
|
||||
*
|
||||
* Applies the options to the parser context. The following options
|
||||
* are never cleared and can only be enabled:
|
||||
*
|
||||
* XML_PARSE_NOERROR
|
||||
* XML_PARSE_NOWARNING
|
||||
* XML_PARSE_NONET
|
||||
* XML_PARSE_NSCLEAN
|
||||
* XML_PARSE_NOCDATA
|
||||
* XML_PARSE_COMPACT
|
||||
* XML_PARSE_OLD10
|
||||
* XML_PARSE_HUGE
|
||||
* XML_PARSE_OLDSAX
|
||||
* XML_PARSE_IGNORE_ENC
|
||||
* XML_PARSE_BIG_LINES
|
||||
*
|
||||
* Returns 0 in case of success, the set of unknown or unimplemented options
|
||||
* in case of error.
|
||||
@ -13355,118 +13599,24 @@ xmlCtxtResetPush(xmlParserCtxtPtr ctxt, const char *chunk,
|
||||
int
|
||||
xmlCtxtUseOptions(xmlParserCtxtPtr ctxt, int options)
|
||||
{
|
||||
if (ctxt == NULL)
|
||||
return(-1);
|
||||
int keepMask;
|
||||
|
||||
if (options & XML_PARSE_RECOVER) {
|
||||
ctxt->recovery = 1;
|
||||
options -= XML_PARSE_RECOVER;
|
||||
ctxt->options |= XML_PARSE_RECOVER;
|
||||
} else
|
||||
ctxt->recovery = 0;
|
||||
if (options & XML_PARSE_DTDLOAD) {
|
||||
ctxt->loadsubset = XML_DETECT_IDS;
|
||||
options -= XML_PARSE_DTDLOAD;
|
||||
ctxt->options |= XML_PARSE_DTDLOAD;
|
||||
} else
|
||||
ctxt->loadsubset = 0;
|
||||
if (options & XML_PARSE_DTDATTR) {
|
||||
ctxt->loadsubset |= XML_COMPLETE_ATTRS;
|
||||
options -= XML_PARSE_DTDATTR;
|
||||
ctxt->options |= XML_PARSE_DTDATTR;
|
||||
}
|
||||
if (options & XML_PARSE_NOENT) {
|
||||
ctxt->replaceEntities = 1;
|
||||
/* ctxt->loadsubset |= XML_DETECT_IDS; */
|
||||
options -= XML_PARSE_NOENT;
|
||||
ctxt->options |= XML_PARSE_NOENT;
|
||||
} else
|
||||
ctxt->replaceEntities = 0;
|
||||
if (options & XML_PARSE_PEDANTIC) {
|
||||
ctxt->pedantic = 1;
|
||||
options -= XML_PARSE_PEDANTIC;
|
||||
ctxt->options |= XML_PARSE_PEDANTIC;
|
||||
} else
|
||||
ctxt->pedantic = 0;
|
||||
if (options & XML_PARSE_NOBLANKS) {
|
||||
ctxt->keepBlanks = 0;
|
||||
ctxt->sax->ignorableWhitespace = xmlSAX2IgnorableWhitespace;
|
||||
options -= XML_PARSE_NOBLANKS;
|
||||
ctxt->options |= XML_PARSE_NOBLANKS;
|
||||
} else
|
||||
ctxt->keepBlanks = 1;
|
||||
if (options & XML_PARSE_DTDVALID) {
|
||||
ctxt->validate = 1;
|
||||
options -= XML_PARSE_DTDVALID;
|
||||
ctxt->options |= XML_PARSE_DTDVALID;
|
||||
} else
|
||||
ctxt->validate = 0;
|
||||
if (options & XML_PARSE_NOWARNING) {
|
||||
options -= XML_PARSE_NOWARNING;
|
||||
ctxt->options |= XML_PARSE_NOWARNING;
|
||||
}
|
||||
if (options & XML_PARSE_NOERROR) {
|
||||
options -= XML_PARSE_NOERROR;
|
||||
ctxt->options |= XML_PARSE_NOERROR;
|
||||
}
|
||||
#ifdef LIBXML_SAX1_ENABLED
|
||||
if (options & XML_PARSE_SAX1) {
|
||||
options -= XML_PARSE_SAX1;
|
||||
ctxt->options |= XML_PARSE_SAX1;
|
||||
}
|
||||
#endif /* LIBXML_SAX1_ENABLED */
|
||||
if (options & XML_PARSE_NODICT) {
|
||||
ctxt->dictNames = 0;
|
||||
options -= XML_PARSE_NODICT;
|
||||
ctxt->options |= XML_PARSE_NODICT;
|
||||
} else {
|
||||
ctxt->dictNames = 1;
|
||||
}
|
||||
if (options & XML_PARSE_NOCDATA) {
|
||||
ctxt->sax->cdataBlock = NULL;
|
||||
options -= XML_PARSE_NOCDATA;
|
||||
ctxt->options |= XML_PARSE_NOCDATA;
|
||||
}
|
||||
if (options & XML_PARSE_NSCLEAN) {
|
||||
ctxt->options |= XML_PARSE_NSCLEAN;
|
||||
options -= XML_PARSE_NSCLEAN;
|
||||
}
|
||||
if (options & XML_PARSE_NONET) {
|
||||
ctxt->options |= XML_PARSE_NONET;
|
||||
options -= XML_PARSE_NONET;
|
||||
}
|
||||
if (options & XML_PARSE_COMPACT) {
|
||||
ctxt->options |= XML_PARSE_COMPACT;
|
||||
options -= XML_PARSE_COMPACT;
|
||||
}
|
||||
if (options & XML_PARSE_OLD10) {
|
||||
ctxt->options |= XML_PARSE_OLD10;
|
||||
options -= XML_PARSE_OLD10;
|
||||
}
|
||||
if (options & XML_PARSE_NOBASEFIX) {
|
||||
ctxt->options |= XML_PARSE_NOBASEFIX;
|
||||
options -= XML_PARSE_NOBASEFIX;
|
||||
}
|
||||
if (options & XML_PARSE_HUGE) {
|
||||
ctxt->options |= XML_PARSE_HUGE;
|
||||
options -= XML_PARSE_HUGE;
|
||||
if (ctxt->dict != NULL)
|
||||
xmlDictSetLimit(ctxt->dict, 0);
|
||||
}
|
||||
if (options & XML_PARSE_OLDSAX) {
|
||||
ctxt->options |= XML_PARSE_OLDSAX;
|
||||
options -= XML_PARSE_OLDSAX;
|
||||
}
|
||||
if (options & XML_PARSE_IGNORE_ENC) {
|
||||
ctxt->options |= XML_PARSE_IGNORE_ENC;
|
||||
options -= XML_PARSE_IGNORE_ENC;
|
||||
}
|
||||
if (options & XML_PARSE_BIG_LINES) {
|
||||
ctxt->options |= XML_PARSE_BIG_LINES;
|
||||
options -= XML_PARSE_BIG_LINES;
|
||||
}
|
||||
ctxt->linenumbers = 1;
|
||||
return (options);
|
||||
/*
|
||||
* For historic reasons, some options can only be enabled.
|
||||
*/
|
||||
keepMask = XML_PARSE_NOERROR |
|
||||
XML_PARSE_NOWARNING |
|
||||
XML_PARSE_NONET |
|
||||
XML_PARSE_NSCLEAN |
|
||||
XML_PARSE_NOCDATA |
|
||||
XML_PARSE_COMPACT |
|
||||
XML_PARSE_OLD10 |
|
||||
XML_PARSE_HUGE |
|
||||
XML_PARSE_OLDSAX |
|
||||
XML_PARSE_IGNORE_ENC |
|
||||
XML_PARSE_BIG_LINES;
|
||||
|
||||
return(xmlCtxtSetOptionsInternal(ctxt, options, keepMask));
|
||||
}
|
||||
|
||||
/**
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user