Since I had no idea of where the problem could come from, I tried to simplify it a bit, and removed the static library that we use (*) so that I can switch back to regular ar and ranlib.
But the bug did not disapear in the move, it still occured on powerpc this time: https://buildd.debian.org/status/fetch.php?pkg=simgrid&arch=powerpc&ver=3.11.1-6&stamp=1404081303 I think I'll disable LTO on all architectures but AMD to work around the bug, so that the package can transition correctly. Thanks for you assistance, Mt (*) instead, the content of our previous static library is recompiled each time it's needed, for each of our 120 our test cases. Ugly and inefficient but I wanted to see. -- Alvin: Sinon, le polymorphisme en C, c'est trop bô. :) Emptty: Ca, c'est clair. Le C, j'aime ca. C'est un peu de l'art primitif, mais ca te secoue les tripes...
signature.asc
Description: Digital signature
