Hi Riccardo, >> Note : except for SimpleAgenda I'm using gnustep packages from Ubuntu 24.04.2 >> (mainly because compiling gnustep is painful) so potential bugs could come >> from those packages and not from upstream gnustep code. In fact, probability >> of a gnustep bug is very low, I'm probably doing something silly... > > compiling on debian with gcc is a breeze... but no worries, using > gnustep packages is fine.
Last time I tried using clang and libobjc2, failed and got bored... I should try again I guess ! > I don't know if there are specific fixes. There was work by Richard and > me to get libxml working with newer libraries, but not of direct bugs. > But Richard constantly fixes things. > If SimpleAgenda code youa re testing is available and the calendar too, > I can try to reproduce for you. Easy enough to reproduce : grab latest code from SimpleAgenda github, run baikal (https://sabre.io/baikal/docker-install/) as a local CalDAV server, add a GroupDAV (probably misnamed) agenda in SimpleAgenda pointing to the baikal server, add an event to this agenda and boom. > the stacktrace is strange, outside the scope of GS. Does it work with > other data? Is it sepcific to that XML fragment? I haven't tested with other data (no other server available) but the XML data seems simple enough, it should not cause problems. Reading the stacktrace I thought the thing was stuck in the depths of gnustep-base (or libobjc), waiting on a lock but I may be wrong. > did you inspect resource and [resource data]? try to spit it out as a > string and see if is complete and valid. I did but will do again. > In case you could get the data, write it to a file (not even needing a > string conversion). > > [[resource data] writeToFile: @"/tmp/testFile" atomically:NO]; > > is your friend. Inspect the file. > > Then write a small test program in the style of: > > NSData *someData = [NSData dataWithContentsOfFile:@"/tmp/testFile"]; > parser = [GSXMLParser parserWithData:someData]; > > and see if it works, files and where it hangs Yep, good idea, I'll try to build a simple reproducer. Thanks, Philippe
