tests: posix + libc/thrd: one tier0 integration platform per arch, improved libc coverage #76377
Conversation
cfriedt
force-pushed
the
tests-posix-common-integration-platforms
branch
4 times, most recently
from
9fbc0da to
fa87c8a
Compare
cfriedt
changed the title
cfriedt
changed the title
zephyrbot
added
area: POSIX
| tests: | ||
| portability.posix.eventfd: {} |
There was a problem hiding this comment.
Given that picolibc is usually the default, it looks like this configuration will test against picolibc twice and never test against minimal libc?
There was a problem hiding this comment.
It's been a while, but IIRC the reason why the base case exists and is empty is to allow people to override things easier on the command line. So, for example, if a completely external libc was used, people could still run the suite without having to n-select anything in addition to y-selecting things.
There was a problem hiding this comment.
I could add an explicit case for the minimal libc though. Good catch 👍
There was a problem hiding this comment.
Thanks, that makes good sense, it's nice to provide a way for people to test things outside of the "known" C libraries. I was mostly wondering if we want to test minimal libc at all, because aiui, it's not happening with the proposed change (I don't know about the pre-patch status).
There was a problem hiding this comment.
Went through all of them and made sure there was a .minimal case as well. So far so good on my desk, so I'll push.
INFO - Added initial list of jobs to queue
INFO - Total complete: 1822/2488 73% skipped: 1341, failed: 0, error: 0
|
If it makes sense to reduce coverage in certain areas, please make suggestions. There is unfortunately a bit of a compromise to make between fast CI and sufficient coverage to ensure things are regression-free, while simultaneously testing most / all architectures and most / all libc's. |
cfriedt
force-pushed
the
tests-posix-common-integration-platforms
branch
from
fa87c8a to
dd4e5fc
Compare
Add one tier0 platform per supported architecture. Signed-off-by: Chris Friedt <[email protected]>
cfriedt
force-pushed
the
tests-posix-common-integration-platforms
branch
from
dd4e5fc to
a34026d
Compare
|
@keith-packard - feel up for a re-review? @nashif - PTAL. Your change request was made. |
|
re-ping @nashif |
The intent of this change is to mitigate the possibility of regressions in POSIX and related areas due to lack of test coverage between multiple architectures and multiple C libraries. There is unfortunately a compromise to make between fast CI and sufficient test coverage.
If it makes sense to reduce the scope in certain areas, please feel free to make suggestions.
Add one tier0 integration platform per supported architecture, and increase libc coverage.
Note
Depending on other suite constraints, only certain tier0 platforms may be used for certain suites (e.g.
netifdependency)