This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 0/N] test-suite improvement - PASS/FAIL: initial patch
On Mon, 17 Sep 2012 14:28:26 +0000, Joseph wrote:
> Also, some tests are shell scripts and may not readily be convertible
> to use test-skeleton.c.
OK, I did not know this. test-skeleton is not the best option then.
> Also, UNSUPPORTED tests may never build or run anything with
> test-skeleton.c at all and so that status needs to be generated from
> the makefiles in such cases.
Right, agreed.
> (I do think every test that is run should create a .out file with the
> stdout/stderr of the test - and that special cases not doing so
> should be fixed to so so.)
The 'grand plan' when I had started working on (and then abandoned
mostly due to distractions) the testsuite a while back was something on
these lines:
1) Isolate the test case run environment. I had added --inhibit-cache
for that reason, but that project kinda fell away.
2) Silence the log on stdout and instead, print out the commands into a
file
3) Add an echo at the end of the target of each test to indicate
PASS/FAIL/WARN, etc.
One of the things I had noted during that time was that each line of a
recipe is invoked in a different shell, so I don't know if the return
codes can be captured as they are being done in this patch -- I assume
that this has been tested though. The way to work around this was to
use the .ONESHELL target. It's not used in our makefiles currently and
I vaguely remember Roland saying that we want to limit our usage of
obscure make features.
Regards,
Siddhesh