On Wed, Dec 31, 2025 at 02:03:45PM -0100, Graham Inggs wrote: > Control: retitle -1 deal.ii: FTBFS on ppc64el: tests segfault > > On Mon, 24 Nov 2025 at 18:09, Adrian Bunk <[email protected]> wrote: > > This bug report mixes two issue: > > > > 1. A recent bug in petsc affecting all architectures, this is now fixed. > > > > 2. Older ppc64el-only test failures: > > https://buildd.debian.org/status/logs.php?pkg=deal.ii&arch=ppc64el > > https://buildd.debian.org/status/fetch.php?pkg=deal.ii&arch=ppc64el&ver=9.6.2-4%2Bb4&stamp=1764009042&raw=0 > > Indeed. > > Looking at the recent FTBFS on ppc64el [1], it was only the "quick" > failures (5-7min) on 2025-11-23 and 2025-11-24 that failed as > described in this bug report. The other failures seem to be due to > segmentation faults during the tests (output copied below). > > > The timing of the start of the ppc64el-only failure would match GCC 15 > > and the ppc64el ABI change, but it is just a guess that this might be > > related. > > I suspect the failure may be related to using the mold linker.
Seems you were right with that: https://buildd.debian.org/status/logs.php?pkg=deal.ii&arch=ppc64el If anyone wants to debug whether it's GCC 15 or the ABI change that is the problem in combination with mold, recent GCC 14 packages also have the ABI change. > Regards > Graham >... cu Adrian
