You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-users@xerces.apache.org by cpplove <sa...@hotmail.com> on 2009/03/24 09:57:18 UTC
MemBufInputSource "Expected comment or CDATA"
Hello everyone,
Im trying to parse wikipedia XML with SAX2 parser of xerces-c 2_7_0.
When i set the parser to parse from MemBufInputSource I get error like
All the files in the collection come up with the same error.
Fatal Error at file R:\000\41000.xml, line 2, char 2
Message: Expected comment or CDATA
But when I try
parser->parse(filename) it does not give me any errors.
my prolem is im having to extend code that already reads the file in a
buffer.
so i want to use memort buffer so i can avoid disk IO again.
it looks like the following
filename = disk.get_first_filename(argv[param]);
file = (unsigned char *)disk.read_entire_file();
//
// read_entire_file() looks like the following
//if ((fp = fopen(filename, "rb")) == NULL) return NULL;
//if (fstat(fileno(fp), &details) == 0)
// if ((*file_length = details.st_size) != 0)
// if ((block = new (std::nothrow) char [(long)(details.st_size + 1)]) !=
NULL)
// +1 for the '\0' on the end
// if (fread(block, (long)details.st_size, 1, fp) == 1)
// block[details.st_size] = '\0';
//
============================================================================
unsigned int zzz = strlen( (const char*)file); // verified this
MemBufInputSource* memIS = new MemBufInputSource (
(const XMLByte*) file ,
zzz ,
filename ,
false );
// memIS->setEncoding(XMLUni::fgUTF8EncodingString); // tried turning this
on
// the file being load is UTF-8
// saved as UTF-8 encoding
parser2->parse(*memIS); // this fails and my handler's events are
never called
parser2->parse(filename); // this works but i want to avoid this in
favor
// of parsing from the buffer.
I tried searching for help but no luck as of yet. I would be very greatful
if you can provide me some insight on what is wrong here.
Best regards
--
View this message in context: http://www.nabble.com/MemBufInputSource--%22Expected-comment-or-CDATA%22-tp22676266p22676266.html
Sent from the Xerces - C - Users mailing list archive at Nabble.com.
Re: MemBufInputSource "Expected comment or CDATA"
Posted by David Bertoni <db...@apache.org>.
cpplove wrote:
> Hello everyone,
>
> Im trying to parse wikipedia XML with SAX2 parser of xerces-c 2_7_0.
>
> When i set the parser to parse from MemBufInputSource I get error like
> All the files in the collection come up with the same error.
>
> Fatal Error at file R:\000\41000.xml, line 2, char 2
> Message: Expected comment or CDATA
This error occurs when the parser encounters the character sequence '<'
'!' which can be either the start of a comment or a CDATA section, but
it finds neither of those constructs. Are you sure the data stream
you're providing is correct, and the length correct? It might be
worthwhile for you to dump the buffer before you parse it, just to see
what the parser is going to get.
>
> But when I try
> parser->parse(filename) it does not give me any errors.
>
> my prolem is im having to extend code that already reads the file in a
> buffer.
> so i want to use memort buffer so i can avoid disk IO again.
> it looks like the following
>
>
> filename = disk.get_first_filename(argv[param]);
> file = (unsigned char *)disk.read_entire_file();
>
> //
> // read_entire_file() looks like the following
> //if ((fp = fopen(filename, "rb")) == NULL) return NULL;
> //if (fstat(fileno(fp), &details) == 0)
> // if ((*file_length = details.st_size) != 0)
> // if ((block = new (std::nothrow) char [(long)(details.st_size + 1)]) !=
> NULL)
>
> // +1 for the '\0' on the end
> // if (fread(block, (long)details.st_size, 1, fp) == 1)
> // block[details.st_size] = '\0';
> //
> ============================================================================
>
> unsigned int zzz = strlen( (const char*)file); // verified this
Can you give more details on what you mean by "verified this?"
> MemBufInputSource* memIS = new MemBufInputSource (
> (const XMLByte*) file ,
> zzz ,
> filename ,
> false );
>
> // memIS->setEncoding(XMLUni::fgUTF8EncodingString); // tried turning this
> on
> // the file being load is UTF-8
> // saved as UTF-8 encoding
This shouldn't be necessary if the stream contains a well-formed document.
Dave
Re: MemBufInputSource "Expected comment or CDATA"
Posted by Ben Griffin <be...@redsnapper.net>.
I Use a DOMLSInput as following:
(here 'parser' is a DOMLSParser made from createLSParser )
DOMDocument* loadDoc(const std::string& xmlfile) {
DOMDocument* rslt = NULL;
if ( ! xmlfile.empty() ) {
DOMLSInput* input = ((DOMImplementationLS*)impl)->createLSInput();
XMLByte* xmlraw = (XMLByte*)(xmlfile.c_str());
MemBufInputSource* mbis = new
MemBufInputSource(xmlraw,xmlfile.size(),X("myfile.xml"), true);
input->setByteStream(mbis);
input->setEncoding(XMLUni::fgUTF8EncodingString); //This must be done.
try {
rslt = parser->parse(input);
}
catch ( ... ) {
//handle errors here..
}
}
return rslt;
}
As far as I understand it, the following should also work, (but
doesn't seem to do so reliably)
DOMDocument* loadDoc(const std::string& xmlfile) {
DOMDocument* rslt = NULL;
if ( ! xmlfile.empty() ) {
DOMLSInput* input = ((DOMImplementationLS*)impl)->createLSInput();
input->setStringData((XMLCh*)xmlfile.c_str()); //very naughty cast,
but we set the encoding next....
input->setEncoding(XMLUni::fgUTF8EncodingString); //This must be done.
try {
rslt = parser->parse(input);
}
catch ( ... ) {
//handle errors here..
}
}
return rslt;
}
I am not sure about what memory needs freeing/releasing - I am not
good at xercesc memory management.
Hope that helps.
-b.
On 24 Mar 2009, at 08:57, cpplove wrote:
unsigned int zzz = strlen( (const char*)file); // verified this
MemBufInputSource* memIS = new MemBufInputSource (
(const XMLByte*) file ,
zzz ,
filename ,
false );
// memIS->setEncoding(XMLUni::fgUTF8EncodingString); // tried turning
this
on
// the file being load is UTF-8
// saved as UTF-8 encoding
parser2->parse(*memIS); // this fails and my handler's events are
never called
parser2->parse(filename); // this works but i want to avoid this
in
favor
// of parsing from the
buffer.