Jan Wielemaker

Yes, but it is in general poor design to rely on the order. Reloading files during development may swap the order. If order matters, consider a priority or dependency.

This is the case for most library foreign predicates processing options. Probably we need to expose the core system option processing to the foreign API.

That is definitely not the final solution :) Have a look at packages/ssl/CMakeList.txt, where you should find a lot of inspiration to do the right check. Note that while working...

As is, we do not build the Windows binaries against tcmalloc. Possibly we should. I do not know how hard it is and what the risks are to replace Windows...

Pushed b8ad01c1abcd9da8699f10f44cc237f73643a3d9, which should fix at least some of this. Wouter's code to query the API to find versions was already in place. This fix avoids interpreting a final digit...

This issue has been mentioned on **SWI-Prolog**. There might be relevant details there: https://swi-prolog.discourse.group/t/graphical-debugger-call-stack-pane-significance-of-arrow-sides-and-gaps-in-the-stack/3232/8

Thanks. This does reproduce. Not yet sure there is a reasonable solution. It gets into a deadlock on the global xpce lock :cry: Unlike SWI-Prolog itself, the graphics subsystem is...

These issues are hard. If I'm not mistaken, cygwin generated code cannot be used in applications using the MS C runtime libs. May also be a calling convention issue. You...

The C stack trace doesn't match up with the Prolog one. I guess Prolog was compiled without debug symbols or the symbols were stripped? So yes, it is most likely...

I think that by default doesn't install the symbols :cry: You'll have to build from source. The `-DCMAKE_BUILD_TYPE=PGO` version does add debug symbols (keeping optimization). I guess you can also...