|
From: David K. <dav...@al...> - 2003-09-24 14:54:59
|
Dejan Krsmanovic wrote: > Hi David, Hi Dejan! > I have applied your patch (I hope it works correctly!). Thanks! It's chugging along happily here, albeit not with very heavy traffic, but nonetheless. I'll shout loudly if I discover any problems related to these changes, of course, but they are rather localized, so hopefully... > Regarding to your question about using threadpool pipeline processor I have > to say that this is still new ferature in development phase so there could > be such errors. Anyway, stack trace you sent looks like error in particular > stage (extract_fid) so it maybe some configuration issue... > I guess Bruce could help you more on this. Well, the stage works fine when using the sync processor, and it is very simple indeed...the config looks like this: # XPath Extract - save the original file_id attribute extract_fid.stageType=XpathExtract extract_fid.nextStage=file_writer extract_fid.XPath.orig_file_id=/newsitem/@file_id I have only looked briefly at the source of the threadpool processor so far, so I don't really have any informed idea about what could be going wrong here. I will look further at both my code, the pipeline configuration and the threadpool source to see if I can find any clues... > Anyway, if you find some Babeldoc's bug please send us a patch. Or we could > grant you CVS access permissions if you are interested! I will happily send patches for suitable scrutiny. I don't think I dare meddle directly with the CVS yet, until I feel a bit more comfortable with the overall codebase of Babeldoc. I hate making changes that cause unexpected regressions elsewhere in the code, and even moreso when it's someone else's. :-) Maybe a bit later? Anyway, thanks much for applying the change! Certainly makes it easier to keep synchronized with the Babeldoc core in CVS. I'll try to clean up the code for my (two, so far) pipeline stages as well, and submit them in a couple of weeks, for consideration. > Dejan David |