Hi On 25-08-16 09:29, Graham Inggs wrote: > I also understood that many of the executables (if not all?) that were > built with the broken version of fpc would be broken.
Well, at least pasdoc isn't broken. I just tried it and it works. > So fpcmkcfg-3.0.0 is just one of many that will crash, and avoiding > the crash there is not fixing the problem. True, but it may help to fix the issue in a broader way once we know the solution and it may help with all the dependent packages that now are blocked on powerpc for migration (release team is considering dropping powerpc as release arch if I remember correctly). I have a fpc package available on debomatic¹ that only for powerpc goes back to the old version of fpcmkcf and I confirm that that works. So I agree with Graham, that this probably just hides the issue that is to be also found elsewhere, at least it would make experimenting easier again if we bootstrap powerpc this way. What do you think? @Gianfranco, your debugging looks awsome. Just a note, this is the final error message of winff when I start in on the porterbox: Inconsistency detected by ld.so: dl-close.c: 811: _dl_close: Assertion `map->l_init_called' failed! pointing again at a similar issue as in the redhat issues. But why only on powerpc. What is so special there? Paul ¹ http://debomatic-powerpc.debian.net/distribution#stable/fpc/3.0.0+dfsg-7~debo2
signature.asc
Description: OpenPGP digital signature
