Using MemBufInputSource with DOM parser

Discussion:

Dan Arcari

2008-05-12 18:05:18 UTC

Hello,

We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource (or
derivation) via our library to xerces so we can parse XML which does not
reside in a file. Is this possible, and could anyone share some examples or
advice on how to go about this?

Thanks

Jesse Pelton

2008-05-12 19:22:49 UTC

Permalink

Take a look at the MemParse sample, which uses a MemBufInputSource to
parse an in-memory document.

-----Original Message-----
From: Dan Arcari [mailto:***@gmail.com]
Sent: Monday, May 12, 2008 2:05 PM
To: c-***@xerces.apache.org
Subject: Using MemBufInputSource with DOM parser

Hello,

We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource (or
derivation) via our library to xerces so we can parse XML which does not
reside in a file. Is this possible, and could anyone share some examples
or
advice on how to go about this?

Thanks

Sahoglu, Ozgur

2008-05-12 19:48:11 UTC

Permalink

Hi Dan,

Parsing a file is no different than parsing a memory buffer (or any
other input source). They all use BinInputStream internally if you look
at the Xerces-C++ implementation. All you need to do is create the
MemBufInputSource and pass it to parser's parse() method (see the API
doc
http://xerces.apache.org/xerces-c/apiDocs/classAbstractDOMParser.html#b6
fbd6cdf84cce77db1c780017d107dd).

DOM parser's parse() method is overloaded to accept a system id (a file
path or URL) or an InputSource, which can be a LocalFileInputSource, a
URLInputSource, a StdInInputSource, a Wrapper4DOMInputSource, or a
MemBufInputSource (see
http://xerces.apache.org/xerces-c/apiDocs/classInputSource.html).

Cheers,

-Ozgur Sahoglu

-----Original Message-----
From: Dan Arcari [mailto:***@gmail.com]
Sent: Monday, May 12, 2008 11:05 AM
To: c-***@xerces.apache.org
Subject: Using MemBufInputSource with DOM parser

Hello,

We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource (or
derivation) via our library to xerces so we can parse XML which does not
reside in a file. Is this possible, and could anyone share some examples
or
advice on how to go about this?

Thanks

Dan Arcari

2008-05-12 23:42:08 UTC

Permalink

Thanks Ozgur.

I have attempted to do the following:

static const char* someXML = "<?xml version="\1.0\" encoding=\"ascii\"
?><message>Hello World</message>";
XMLCh* testData = XMLString::transcode(someXML);
XMLByte* pXMLByte = reinterpret_cast<XMLByte*>(testData);

MemBufInputSource* memIS = new MemBufInputSource(pXMLByte, strlen(someXML),
"test", false);

mParser->parse((const DOMInputSource&)memIS);

---
//mParser is created as:
mParser =
((DOMImplementationLS&)impl)->createDOMBuilder(DOMImplementationLS::MODE_SYNCHRONOUS,
0);

Using this code, I receive a segmentation fault in ReaderMgr where it
attempts to call src.makeStream().

Do I need to prepare my memIS object some other way so that it will properly
provide a BinInputStream object to the functions farther down the line?

Thanks,

Dan

Post by Sahoglu, Ozgur
Hi Dan,
Parsing a file is no different than parsing a memory buffer (or any
other input source). They all use BinInputStream internally if you look
at the Xerces-C++ implementation. All you need to do is create the
MemBufInputSource and pass it to parser's parse() method (see the API
doc
http://xerces.apache.org/xerces-c/apiDocs/classAbstractDOMParser.html#b6
fbd6cdf84cce77db1c780017d107dd<http://xerces.apache.org/xerces-c/apiDocs/classAbstractDOMParser.html#b6fbd6cdf84cce77db1c780017d107dd>
).
DOM parser's parse() method is overloaded to accept a system id (a file
path or URL) or an InputSource, which can be a LocalFileInputSource, a
URLInputSource, a StdInInputSource, a Wrapper4DOMInputSource, or a
MemBufInputSource (see
http://xerces.apache.org/xerces-c/apiDocs/classInputSource.html).
Cheers,
-Ozgur Sahoglu
-----Original Message-----
Sent: Monday, May 12, 2008 11:05 AM
Subject: Using MemBufInputSource with DOM parser
Hello,
We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource (or
derivation) via our library to xerces so we can parse XML which does not
reside in a file. Is this possible, and could anyone share some examples or
advice on how to go about this?
Thanks

Sahoglu, Ozgur

2008-05-12 23:59:30 UTC

Permalink

Dan,

Remove the cast in the parse() method. MemBufInputSource is-a
InputSource. Use this code instead:

MemBufInputSource* memIS = new MemBufInputSource((const XMLByte
*)someXML, strlen(someXML), "test", false);

mParser->parse(*memIS);

Cheers,

-Ozgur Sahoglu

-----Original Message-----
From: Dan Arcari [mailto:***@gmail.com]
Sent: Monday, May 12, 2008 4:42 PM
To: c-***@xerces.apache.org
Subject: Re: Using MemBufInputSource with DOM parser

Thanks Ozgur.

I have attempted to do the following:

static const char* someXML = "<?xml version="\1.0\" encoding=\"ascii\"
?><message>Hello World</message>";
XMLCh* testData = XMLString::transcode(someXML);
XMLByte* pXMLByte = reinterpret_cast<XMLByte*>(testData);

MemBufInputSource* memIS = new MemBufInputSource(pXMLByte,
strlen(someXML),
"test", false);

mParser->parse((const DOMInputSource&)memIS);

---
//mParser is created as:
mParser =
((DOMImplementationLS&)impl)->createDOMBuilder(DOMImplementationLS::MODE
_SYNCHRONOUS,
0);

Using this code, I receive a segmentation fault in ReaderMgr where it
attempts to call src.makeStream().

Do I need to prepare my memIS object some other way so that it will
properly
provide a BinInputStream object to the functions farther down the line?

Thanks,

Dan

On Mon, May 12, 2008 at 12:48 PM, Sahoglu, Ozgur

http://xerces.apache.org/xerces-c/apiDocs/classAbstractDOMParser.html#b6
fbd6cdf84cce77db1c780017d107dd<http://xerces.apache.org/xerces-c/apiDocs
/classAbstractDOMParser.html#b6fbd6cdf84cce77db1c780017d107dd>

Post by Sahoglu, Ozgur
).
DOM parser's parse() method is overloaded to accept a system id (a file
path or URL) or an InputSource, which can be a LocalFileInputSource, a
URLInputSource, a StdInInputSource, a Wrapper4DOMInputSource, or a
MemBufInputSource (see
http://xerces.apache.org/xerces-c/apiDocs/classInputSource.html).
Cheers,
-Ozgur Sahoglu
-----Original Message-----
Sent: Monday, May 12, 2008 11:05 AM
Subject: Using MemBufInputSource with DOM parser
Hello,
We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource (or
derivation) via our library to xerces so we can parse XML which does not
reside in a file. Is this possible, and could anyone share some examples
or
advice on how to go about this?
Thanks

Alberto Massari

2008-05-13 00:17:41 UTC

Permalink

Wrapper4InputSource wrap(memIS, false);
mParser->parse(wrap);

Alberto

Post by Sahoglu, Ozgur
Dan,
Remove the cast in the parse() method. MemBufInputSource is-a
MemBufInputSource* memIS = new MemBufInputSource((const XMLByte
*)someXML, strlen(someXML), "test", false);
mParser->parse(*memIS);
Cheers,
-Ozgur Sahoglu
-----Original Message-----
Sent: Monday, May 12, 2008 4:42 PM
Subject: Re: Using MemBufInputSource with DOM parser
Thanks Ozgur.
static const char* someXML = "<?xml version="\1.0\" encoding=\"ascii\"
?><message>Hello World</message>";
XMLCh* testData = XMLString::transcode(someXML);
XMLByte* pXMLByte = reinterpret_cast<XMLByte*>(testData);
MemBufInputSource* memIS = new MemBufInputSource(pXMLByte,
strlen(someXML),
"test", false);
mParser->parse((const DOMInputSource&)memIS);
---
mParser =
((DOMImplementationLS&)impl)->createDOMBuilder(DOMImplementationLS::MODE
_SYNCHRONOUS,
0);
Using this code, I receive a segmentation fault in ReaderMgr where it
attempts to call src.makeStream().
Do I need to prepare my memIS object some other way so that it will properly
provide a BinInputStream object to the functions farther down the line?
Thanks,
Dan
On Mon, May 12, 2008 at 12:48 PM, Sahoglu, Ozgur

Post by Sahoglu, Ozgur
Hi Dan,
Parsing a file is no different than parsing a memory buffer (or any
other input source). They all use BinInputStream internally if you

look

Post by Sahoglu, Ozgur
at the Xerces-C++ implementation. All you need to do is create the
MemBufInputSource and pass it to parser's parse() method (see the API
doc

Post by Sahoglu, Ozgur
).
DOM parser's parse() method is overloaded to accept a system id (a

file

Post by Sahoglu, Ozgur
path or URL) or an InputSource, which can be a LocalFileInputSource, a
URLInputSource, a StdInInputSource, a Wrapper4DOMInputSource, or a
MemBufInputSource (see
http://xerces.apache.org/xerces-c/apiDocs/classInputSource.html).
Cheers,
-Ozgur Sahoglu
-----Original Message-----
Sent: Monday, May 12, 2008 11:05 AM
Subject: Using MemBufInputSource with DOM parser
Hello,
We have a library written on top of Xerces-c which uses DOM parsing of
file-based XML. We would like to be able to send a MemBufInputSource

(or

Post by Sahoglu, Ozgur
derivation) via our library to xerces so we can parse XML which does

not

Post by Sahoglu, Ozgur
reside in a file. Is this possible, and could anyone share some

examples

Post by Sahoglu, Ozgur
or
advice on how to go about this?
Thanks

Dan Arcari

2008-05-13 00:44:30 UTC

Permalink

Thanks Alberto. This was successful!

Post by Alberto Massari
Wrapper4InputSource wrap(memIS, false);
mParser->parse(wrap);
Alberto

Post by Sahoglu, Ozgur
Dan,
Remove the cast in the parse() method. MemBufInputSource is-a
MemBufInputSource* memIS = new MemBufInputSource((const XMLByte
*)someXML, strlen(someXML), "test", false);
mParser->parse(*memIS);
Cheers,
-Ozgur Sahoglu
-----Original Message-----
4:42 PM
Subject: Re: Using MemBufInputSource with DOM parser
Thanks Ozgur.
static const char* someXML = "<?xml version="\1.0\" encoding=\"ascii\"
?><message>Hello World</message>";
XMLCh* testData = XMLString::transcode(someXML);
XMLByte* pXMLByte = reinterpret_cast<XMLByte*>(testData);
MemBufInputSource* memIS = new MemBufInputSource(pXMLByte,
strlen(someXML),
"test", false);
mParser->parse((const DOMInputSource&)memIS);
---
mParser =
((DOMImplementationLS&)impl)->createDOMBuilder(DOMImplementationLS::MODE
_SYNCHRONOUS,
0);
Using this code, I receive a segmentation fault in ReaderMgr where it
attempts to call src.makeStream().
Do I need to prepare my memIS object some other way so that it will properly
provide a BinInputStream object to the functions farther down the line?
Thanks,
Dan
On Mon, May 12, 2008 at 12:48 PM, Sahoglu, Ozgur

Post by Sahoglu, Ozgur
Hi Dan,
Parsing a file is no different than parsing a memory buffer (or any
other input source). They all use BinInputStream internally if you

look

Post by Sahoglu, Ozgur
at the Xerces-C++ implementation. All you need to do is create the
MemBufInputSource and pass it to parser's parse() method (see the API
doc

http://xerces.apache.org/xerces-c/apiDocs/classAbstractDOMParser.html#b6
fbd6cdf84cce77db1c780017d107dd<
http://xerces.apache.org/xerces-c/apiDocs
/classAbstractDOMParser.html#b6fbd6cdf84cce77db1c780017d107dd>

Post by Sahoglu, Ozgur
).
DOM parser's parse() method is overloaded to accept a system id (a

file

(or

Post by Sahoglu, Ozgur
derivation) via our library to xerces so we can parse XML which does

not

Post by Sahoglu, Ozgur
reside in a file. Is this possible, and could anyone share some

examples

Post by Sahoglu, Ozgur
or
advice on how to go about this?
Thanks