diff options
| author | Paul Eggleton <paul.eggleton@linux.microsoft.com> | 2020-06-14 21:28:45 -0700 |
|---|---|---|
| committer | Khem Raj <raj.khem@gmail.com> | 2020-06-14 22:31:44 -0700 |
| commit | 3251fe210a91d13ab2a6c5b7ecb283aa8e019020 (patch) | |
| tree | 0c30f181d822ee38d9fb8472e19a6081efeb884b /meta-python/recipes-devtools/python/python-incremental.inc | |
| parent | 12072f84377897de9117481c9d173d71f4e0b307 (diff) | |
| download | meta-openembedded-3251fe210a91d13ab2a6c5b7ecb283aa8e019020.tar.gz | |
protobuf-c: disable parallelism to avoid race condition
Unfortunately 489d3b4b932ee8016d792341f8ea5836a9522cd4 did not completely
fix the problem - if you try cleaning and rebuilding protobuf-c-native it
doesn't take long to reproduce the issue on a 32-core machine. I spent
some time trying to debug this but failed, there is still a race between
generating t.test-full.pb.h and compiling cxx_generate_packed_data.c
despite BUILT_SOURCES and explicit dependencies. I even tried converting
the multiple target rules to use grouped targets (&:), that didn't fix it
either. Disabling parallelism as a workaround only costs ~20s and it
turns out that upstream is switching to Meson soon anyway:
https://github.com/protobuf-c/protobuf-c/pull/340
Signed-off-by: Paul Eggleton <paul.eggleton@linux.microsoft.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Diffstat (limited to 'meta-python/recipes-devtools/python/python-incremental.inc')
0 files changed, 0 insertions, 0 deletions