You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-users@xalan.apache.org by Arindam <ar...@gmail.com> on 2007/07/09 16:38:59 UTC
Building Xalan 1.10 with Microsoft Windows Platform SDK 2003
Hi,
I have been trying to build Xalan 1.10 with the MS Windows Platform
SDK 2003 compiler and CRT library. In fact I intend to build it for
X64 (AMD64) - but that seems a distant goal.
For now I am stuck at more basic issues:
1. I have built Xerces (2.7).
2. I want to build Xalan 1.10 and I am faced with the following:
- There are no .mak files for nmake to use in the Xalan-C++ 1.10 distribution.
- I thought of generating nmake files using the Export Makefiles
option of the Visual C++ 6.0 IDE. But for that I need to open the .dsp
files in it first. I did that ... but somehow, the Export Makefile
operation did not seem to generate any output ... either on the screen
or on the disk.
- Some of the parts like ApacheModuleXSLT seems to require basic
Apache development headers and libraries. I am not sure where to pick
them up from.
I want to know how else can I generate the makefiles for nmake to use,
or if they are readily available somewhere. Using VS2005 is not an
option.
I would really appreciate any pointers
Thanks,
Arindam
Re: Ignoring the whitespace in text nodes
Posted by David Bertoni <db...@apache.org>.
Bharadwaj, Anand (ELS-OXF) wrote:
> Hi,
>
> Thanks for the update.
>
> When we run the below Xpath query on an another application (XMLSpy),
> it ignores the first node (containing white spaces)
> and returns the count as 4.
The Microsoft XSLT processor strips all whitespace text nodes from the
source tree before transforming it. Perhaps XMLSpy is relying on that
processor.
>
> We would like to request if it is safe to assume that the mentioned
> application handles this scenario independently and not as part of
> implementation of neither of XML nor XPath specifications.
You will have to ask Altova that question, not me.
Dave
RE: Ignoring the whitespace in text nodes
Posted by "Bharadwaj, Anand (ELS-OXF)" <A....@elsevier.com>.
Hi,
Thanks for the update.
When we run the below Xpath query on an another application (XMLSpy),
it ignores the first node (containing white spaces)
and returns the count as 4.
We would like to request if it is safe to assume that the mentioned
application handles this scenario independently and not as part of
implementation of neither of XML nor XPath specifications.
Thanks and Regards,
Anand
-----Original Message-----
From: David Bertoni [mailto:dbertoni@apache.org]
Sent: 10 July 2007 18:01
To: xalan-c-users@xml.apache.org
Subject: Re: Ignoring the whitespace in text nodes
Bharadwaj, Anand (ELS-OXF) wrote:
> Hi All,
>
> We are having an issue with XSLT parser sample (SimpleXPathAPI.exe).
It's an XPath _processor_, not an XPath _parser_.
>
> When the XPath query (count(/*[1]/*[7]/*[5]/*[1]/*[3]/*[2]/node()))is
> made on the document containing the XML node snippet:
>
> "<para id="para13">
>
> <italic>Expand one's cultural knowledge</italic>
>
...
>
> The parser is considering the leading 'whitespace' characters as one
> node and the total node count as 5 which are,
>
> 1. *whitespace before <italic>*,
>
> 2. <italic>
>
> 3. text between <italic> and <anchor>
>
> 4. <anchor>
>
> 5. text after <anchor> till end.
This is correct, since that's what the XML recommendation, the XPath
data
model, and XML Information Set recommendation require.
>
> As per our requirement the count should be 4 (excluding the (1)
above).
>
> We request to kindly let us know how to ignore these 'whitespace'
nodes
> which are
>
> actually children of 'Text node' types, through Xpath queries since we
> do not have the freedom to reload the document in any other
configuration.
The only thing you can do is to build the document and strip the
whitespace-only text nodes before they get to the XPath processor. You
would need to write your own code to do that, or you could write a
stylesheet to remove them, transform the original document using the
XSLT
processor, then use the transformed document with SimpleXPathAPI.exe.
Dave
Re: Ignoring the whitespace in text nodes
Posted by David Bertoni <db...@apache.org>.
Bharadwaj, Anand (ELS-OXF) wrote:
> Hi All,
>
> We are having an issue with XSLT parser sample (SimpleXPathAPI.exe).
It's an XPath _processor_, not an XPath _parser_.
>
> When the XPath query (count(/*[1]/*[7]/*[5]/*[1]/*[3]/*[2]/node()))is
> made on the document containing the XML node snippet:
>
> "<para id="para13">
>
> <italic>Expand one's cultural knowledge</italic>
>
...
>
> The parser is considering the leading 'whitespace' characters as one
> node and the total node count as 5 which are,
>
> 1. *whitespace before <italic>*,
>
> 2. <italic>
>
> 3. text between <italic> and <anchor>
>
> 4. <anchor>
>
> 5. text after <anchor> till end.
This is correct, since that's what the XML recommendation, the XPath data
model, and XML Information Set recommendation require.
>
> As per our requirement the count should be 4 (excluding the (1) above).
>
> We request to kindly let us know how to ignore these 'whitespace' nodes
> which are
>
> actually children of 'Text node' types, through Xpath queries since we
> do not have the freedom to reload the document in any other configuration.
The only thing you can do is to build the document and strip the
whitespace-only text nodes before they get to the XPath processor. You
would need to write your own code to do that, or you could write a
stylesheet to remove them, transform the original document using the XSLT
processor, then use the transformed document with SimpleXPathAPI.exe.
Dave
Re: Ignoring the whitespace in text nodes
Posted by da...@us.ibm.com.
First, look at the whitespace stripping options in section 3.4 of the XSLT
spec. If that doesn't get what you want, you can apply a predicate to
deselect some nodes when you count them. Perhaps
count(blah/node()[normalize-space()!=' '])
will obtain what you want.
.................David Marston
Ignoring the whitespace in text nodes
Posted by "Bharadwaj, Anand (ELS-OXF)" <A....@elsevier.com>.
Hi All,
We are having an issue with XSLT parser sample (SimpleXPathAPI.exe).
When the XPath query (count(/*[1]/*[7]/*[5]/*[1]/*[3]/*[2]/node()))is
made on the document containing the XML node snippet:
"<para id="para13">
<italic>Expand one's cultural knowledge</italic>
and institutionalize it so that it can be accessed and incorporated into
the delivery of services. We must attempt to seek out sociocultural
information about the individual patient that will then help us have a
better feel for how to perform an interview or history-what to ask, how
to ask-and how to modify treatment interventions appropriately on the
basis of a <anchor id="p27"/>person's cultural reality. It is
impossible, and unnecessary, to learn all there is to know about all
cultural subgroups, but clinicians must be aware of the ethnographic
information related to the local community and relevant beliefs and
behaviors of their patients and the patients' families.</para>"
The parser is considering the leading 'whitespace' characters as one
node and the total node count as 5 which are,
1. whitespace before <italic>,
2. <italic>
3. text between <italic> and <anchor>
4. <anchor>
5. text after <anchor> till end.
As per our requirement the count should be 4 (excluding the (1) above).
We request to kindly let us know how to ignore these 'whitespace' nodes
which are
actually children of 'Text node' types, through Xpath queries since we
do not have the freedom to reload the document in any other
configuration.
Regards,
Anand
Re: Building Xalan 1.10 with Microsoft Windows Platform SDK 2003
Posted by David Bertoni <db...@apache.org>.
Arindam wrote:
> Hi,
>
> I have been trying to build Xalan 1.10 with the MS Windows Platform
> SDK 2003 compiler and CRT library. In fact I intend to build it for
> X64 (AMD64) - but that seems a distant goal.
>
> For now I am stuck at more basic issues:
>
> 1. I have built Xerces (2.7).
> 2. I want to build Xalan 1.10 and I am faced with the following:
>
> - There are no .mak files for nmake to use in the Xalan-C++ 1.10
> distribution.
We should probably implement a way to build executables using GNU make and
the existing makefiles. Unfortunately, MS has decided not to support NMAKE
and auto-generated makefiles any longer.
>
> - I thought of generating nmake files using the Export Makefiles
> option of the Visual C++ 6.0 IDE. But for that I need to open the .dsp
> files in it first. I did that ... but somehow, the Export Makefile
> operation did not seem to generate any output ... either on the screen
> or on the disk.
Are you sure you installed VS correctly, and you checked every project in
the export dialog box? I just tried this with a clean Xalan-C++ 1.10
source tree, and it worked fine. What happens if you check the "Write
dependencies when writing makefiles" checkbox?
>
> - Some of the parts like ApacheModuleXSLT seems to require basic
> Apache development headers and libraries. I am not sure where to pick
> them up from.
You can either get them from Apache, or just ignore that sample
applications. It has been removed in later versions anyway.
> I want to know how else can I generate the makefiles for nmake to use,
> or if they are readily available somewhere. Using VS2005 is not an
> option.
I can't think of any way to do this. Even VS2005 cannot export makefiles,
so you'd have to build using the project files.
Dave