Hi,

On Friday, March 21, 2025 11:27 CET, Riccardo Mottola 
<[email protected]> wrote:

> Hi,
> 
> Sebastian Reitenbach wrote:
> > this is not always reproducible, but I've seen it crashing in WebServices, 
> > but also as below now, in Performance, here's one I properly catched with 
> > some information:
> 
> It could be that some change in the upgrade broke it.
> 
> I tried building several times performance on FreeBSD (clang/libobjc2 
> setup with graphviz) and had no issues.
> I also tried a couple of times with gcc without graphviz on linux, no 
> issues either.

The thing is, I used the ports tree to build these packages, and where the 
issue was observed in bulk package builds.
When I go straght to /usr/ports/x11/gnustep/performance or .../webservices, and 
just built those single packages, then I never was able to trigger the issue.
The issue only appeared, when I go to: /usr/ports/x11/gnustep and do
sudo pkg_delete gnustep-make, gnustep-base     <- which in turn uninstalls all 
GNUstep related packages depending on those
then run a make clean && make install
the make install in the top directory, is also cleans dependencies, i.e. 
removes build sources from base, gui, etc. 
While on the other hand, with those existing, never saw the crash, only when 
they are removed, in that "build all GNUstep" case.

> 
> >
> > Making all for doc Performance...
> > autogsdoc -Project Performance -DocumentationDirectory Performance  
> > -MakeDependencies Performance/dependencies GSCache.h GSFIFO.h 
> > GSIOThreadPool.h GSLinkedList.h GSThreadPool.h GSThroughput.h GSTicker.h 
> > GSSkipMutableArray.h GSUniqued.h
> > 2025-03-19 11:02:28.867 autogsdoc[40804:8011684538824] String mismatch in 
> > merge for (type, NSUInteger). S:GSLinkedList, D:GSIOThreadPool
> > gmake[1]: *** Deleting file 'Performance/dependencies'
> > gmake[1]: *** No rule to make target 'Performance/dependencies', needed by 
> > 'internal-doc-all_'.  Stop.
> > gmake: *** [/usr/local/share/GNUstep/Makefiles/Master/rules.make:297: 
> > Performance.all.doc.variables] Error 2
> 
> 
> IIRC you can disable it with documentation=no

ah, good to know, I didn't find that. as a stopgap measure, I did set 
AUTOGSDOC=echo and updated the ports.


> 
> Riccardo


Reply via email to