Closed ghost closed 3 years ago
Thanks for the PR. I don't think we even end up printing a status if ntotal
is 0. I think the problem is that the initial formatstatus(NULL, 0)
is done to ensure that the status format string is valid, even if we never end up using it.
I pushed a commit to return from build()
early in that case, which should solve the problem. Can you verify that gnome-backgrounds builds successfully with latest master?
Found during a FreeBSD Ports exp-run that uses samurai instead of ninja
I'm curious about the other results from this run. I wouldn't be surprised if there are a handful of argument-order bugs (i.e. samu foo -v
) and maybe some underspecified dependencies, but I'm interested to know if there are any other issues you've hit.
Thanks for fixing this so quickly. Yes, gnome-background builds fine with the latest master. I'll let you know if we run into more problems or how the exp-run went overall.
Samu dies with SIGFPE when NINJA_STATUS="%p [%s/%t]" with a basically empty build file:
This happens for example during the gnome-backgrounds build.
Found during a FreeBSD Ports exp-run that uses samurai instead of ninja:
http://package18.nyi.freebsd.org/data/122amd64-default-foo/2021-04-04_07h37m38s/logs/errors/gnome-backgrounds-3.38.0.log