Project

Profile

Help

Saxon C version

Added by Anonymous about 14 years ago

Legacy ID: #8633613 Legacy Poster: Alex Grosu (alexjavauser)

Hi, I'm working on a project involving distributed C/C++ computing and XSLT2.0 operations on a grid that doesn't support java. I've gave given up hope for the C++ version of Saxon and I hope there still is a way to do it by using plain old C. Are there any Saxon libraries written in C (or ideally C++)? Thank you, Alex


Replies (12)

Please register to reply

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8633643 Legacy Poster: Michael Kay (mhkay)

Sorry, you're out of luck. Saxon source code is 100% Java (except for a few modules that provide a C# API for .NET).

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8633649 Legacy Poster: Alex Grosu (alexjavauser)

I guess I need to look for alternatives. Thank you for your answer. It was very fast. Alex

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8633894 Legacy Poster: Michael Ferrando (tychicus)

Alex, I don't know your situation, but I have been using MLJAM with MarkLogic. I need Michael Kay's Saxon for my stylesheets and extensions. It allows you to write java inside xQuery functions. There is a servlet that uses beanshell to compile the code and then returns. Anyway, I thought it might give you an idea of how to do this in your situation. Sincerely Tychicus http://developer.marklogic.com/learn/2006-05-mljam

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8635848 Legacy Poster: Michael Ferrando (tychicus)

Alex, I also found that there is a C version of Xalan. I don't know how up-to-date it is, but I did find it. -mike

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8635914 Legacy Poster: Alex Grosu (alexjavauser)

Hi Mike, I need to use XLST2.0. The XML translation implies many functions and types specific to the 2.0 version of XSLT. I know of a C version of xalan, but is only compatible with the 1.0 version of XSLT. I googled a lot and always I find the same truth: officially, the XSLT 2.0 is supported only for java. Alex

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8635953 Legacy Poster: Michael Kay (mhkay)

I'm not sure what you mean by "officially". There are quite a few implementations of XSLT 2.0 now, but this isn't the place to discuss them.

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636012 Legacy Poster: David Lee (daldei)

As mentioned by Tychicus, an option ito run XSLT 2.0 transformations using Saxon is by a web service. Tychicus mentioned MLJAM which is MarkLogic specific. xmlsh has a servlet module which is not dependant on MarkLogic, ( But is dependent on tomcat and java). http://www.xmlsh.org/EmbeddingServlet I use this to run saxon xslt2 transformations in a system that is not java based naturally so uses a HTTP call to run the xslt. Another option is to use JNI to call into saxon from your C or C++ program.

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636104 Legacy Poster: Alex Grosu (alexjavauser)

The XSLT transform will apply to a high volume data ( half a million xml files, over 5GB total size) in a window of 30 minutes over a 100Mbps network. Will the HTTP be able to handle it?

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636123 Legacy Poster: David Lee (daldei)

I dont see why HTTP would be the bottleneck, but there may be many other things that are. For volume of that size I'd suggest "rolling your own" servelet so that it could take advantaged of advanced Saxon features such as pre-compiling the stylesheet, and definitely consider Saxon EE which can do some things in streaming mode as well as is more optimized. I suspect your volume requirements will definately be pushing the envelope but maybe it wont burst open :) But again I dont see why HTTP itself would be the bottleneck.

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636145 Legacy Poster: Michael Kay (mhkay)

The XSLT transform will apply to a high volume data ( half a million xml files, over 5GB total size) in a window of 30 minutes over a 100Mbps network. Will the HTTP be able to handle it? I think that workload is not in the range where one could say "absolutely, no problem", nor is it in the range where one can say "not a hope, forget it". It needs detailed sizing.

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636376 Legacy Poster: Michael Ferrando (tychicus)

daldei, I looked at xmlsh. Could I call this from MarkLogic xquery? From reading about xmlsh, I was not sure I could use this with MarkLogic. -mike

RE: Saxon C version - Added by Anonymous about 14 years ago

Legacy ID: #8636471 Legacy Poster: David Lee (daldei)

This is getting a bit off topic (but were still talking about saxon). Yes you can call this from MarkLogic, (via http post) I do it myself. For questions on xmlsh please contact me offline and I can send you some sample code. David Lee

    (1-12/12)

    Please register to reply