On 7/16/20 9:52 PM, Michael Roth wrote: > Quoting Philippe Mathieu-Daudé (2020-07-16 12:59:28) >> On 7/16/20 7:55 PM, Michael Roth wrote: >>> Quoting Peter Maydell (2020-07-16 05:53:17) >>>> The first merge I tried to process after bumping VERSION for rc0 >>>> failed on test-qga like this: >>>> >>>> MALLOC_PERTURB_=${MALLOC_PERTURB_:-$(( ${RANDOM:-0} % 255 + 1))} >>>> tests/test-qga -m=quick -k --tap < /dev/null | ./scripts/tap-driv >>>> er.pl --test-name="test-qga" >>>> PASS 1 test-qga /qga/sync-delimited >>>> PASS 2 test-qga /qga/sync >>>> PASS 3 test-qga /qga/ping >>>> ** >>>> ERROR:/home/petmay01/linaro/qemu-for-merges/tests/test-qga.c:303:test_qga_info: >>>> assertion failed (version == QEMU_VERSION): ("5.0.9 >>>> 0" == "5.0.50") >>>> ERROR test-qga - Bail out! >>>> ERROR:/home/petmay01/linaro/qemu-for-merges/tests/test-qga.c:303:test_qga_info: >>>> assertion failed (versio >>>> n == QEMU_VERSION): ("5.0.90" == "5.0.50") >>>> Aborted (core dumped) >>>> /home/petmay01/linaro/qemu-for-merges/tests/Makefile.include:659: >>>> recipe for target 'check-unit' failed >>>> >>>> Looking at timestamps on files, tests/test-qga.o never got rebuilt, >>>> even though config-host.h has been updated (and so has the new >>>> QEMU_VERSION). Any idea what's gone wrong here? >>>> >>>> Also weird: this build tree has no .d files in it. >>> >>> I've been trying to reproduce with: >>> >>> make >>> make check-unit >>> *bump VERSION >>> make check-unit >>> >>> but test-qga.o gets rebuilt as expected and the test passed. >>> >>> This is with ubuntu 18.04, x86, with out-of-tree build directory. Are you >>> aware >>> of any other factors that might be needed to reproduce this? >> >> The problem is not for qga, it affects all QEMU objects. > > But is it intermittent, environment-dependent? I'm trying to understand how to > replicate Peter's result since it seems like it would be straightforward > reproducer.
How to reproduce: https://www.mail-archive.com/qemu-devel@nongnu.org/msg723531.html > >> >>> >>>> >>>> thanks >>>> -- PMM >>> >> >