[Fuego] LTP unneeded build (was RE: Functional.LTP failed with "./ltp_target_run.sh not found")

Liu, Wenlong liuwl.fnst at cn.fujitsu.com
Wed Nov 15 06:09:06 UTC 2017


> -----Original Message-----
> From: Bird, Timothy [mailto:Tim.Bird at sony.com]
> Sent: Tuesday, November 14, 2017 10:43 AM
> To: Liu, Wenlong/刘 文龙 <liuwl.fnst at cn.fujitsu.com>;
> fuego at lists.linuxfoundation.org
> Subject: LTP unneeded build (was RE: Functional.LTP failed with
> "./ltp_target_run.sh not found")
> 
> > -----Original Message-----
> > From: Liu, Wenlong [mailto:liuwl.fnst at cn.fujitsu.com]
> > A small problem:
> > It seems that LTP will always be built at the first time(if LTP has
> > never been built for the target ARCH before), although I have
> > specified the "homedir" in spec file.
> >
> > I know the default "phases" is "build deploy run", but what I want is
> > just run LTP in the folder I specified and skip "build deploy".
> > (I don't want to add "phases" to spec file at the same time when I
> > specified the "homedir".)
> >
> > I don't know if it was what you want.
> > But for me, LTP "build" should be skipped when I specified "homedir"
> > in spec file.
> > If you think same with me, I'll fix it.
> > If not, I'll specify "phases" to the spec file for my own jobs.
> 
> You're right.  The logic in test_pre_check was not correct.  The build
> phase should be skipped when LTP is detected on the target board.  However,
> we do need to do a deploy phase, as there are a few Fuego-specific files
> that need to be copied over.  We could try to detect if these are already
> there, but they are small enough that the copy operation is fast, and it's
> not a pain to copy them every time we run the test.
> 
> In any event, I made a fix for this, and checked it into my master branch
> (commit aba936a).
> Please test it out.

Oh, yes, the deploy phase should not be skipped.
Thanks for your fix.
I'll test it out and let you know if I have any other questions.

> Thanks for the bug report.
> 
> There is one issue outstanding, that I'm still working on, and that is an
> additional LTP patch that consolidates posix results into a single file.
> Since our posix test parsing is currently non-functional, this patch has
> no effect at the moment.  I'm hoping to change the parsing code for LTP
> posix tests so that the patch is not needed.

Thanks.

>  -- Tim
> 
> 
> 





More information about the Fuego mailing list