On Mon, 23 Jun 2003 [EMAIL PROTECTED] wrote: Hi,
> > The great questions are: > > a) why does mbstowcs(..) fail on 4-STABLE bu not on 5.1 ? > > b) [more important] how to correctly recover from this error path ? > > I do not ocnsider it a FreeBSD specific problem but simply an > > unhandled error path in xalan. > > It's strange that mbstowcs() would fail, but I seem to recall this was a > problem in previous attempts to port Xerces-C to FreeBSD. I found the problem and going to work around it as soon as I can find the time. It got fixed in 4-STABLE 2003/04/05 and is fixed in 5* and HEAD. Up to this time they didn't count if *wcstring == 0 See following commit: src/lib/libc/locale/ansi.c,v 1.3.6.1 2003/04/05 08:28:24 tjr Exp $" http://www.freebsd.org/cgi/cvsweb.cgi/src/lib/libc/locale/Attic/ansi.c.diff?r1=1.3&r2=1.3.6.1&f=h > However, I disagree this is an unhandled error path in Xalan. We should > make XalanTransformer::initialize() exception-safe, but it _should_ > propagate any exception thrown, since there's no clean way to handle this > in the library. Callers must be prepared to handle exceptions from > XalanTransformer::initialize() and we should make sure it's documented that > way. Ok, sorry you are right of course. Might have been irritating becuas my catch(...) was deep in the sources but that simply had been like that for debugging. What I meant was that Xalan(ExE) doesn't catch any exceptions. If this gets documented and all samples get adjusted everything would be fine of course. Ahh and btw. if we are with documentation. Would it be possible to include pre-build documentation in the src tarball with the following releases as xerces-c does. It is very "expensive" to build them in the ports system as from what I can see one needs java and that needs to be fetch manually due to licensing.... Or perhaps have a seperate Xalan-c-1.x-apidocs.tar.gz for this ? -- Greetings Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT 56 69 73 69 74 http://www.zabbadoz.net/
