public inbox for libabigail@sourceware.org
 help / color / mirror / Atom feed
* Re: Where are we?
       [not found] <22b479ce-09a5-52c6-24f9-d959310ce8f2@redhat.com>
@ 2022-09-22  1:34 ` Ben Woodard
       [not found] ` <16c39076-2d42-0583-a013-7c9c0005f841@redhat.com>
  1 sibling, 0 replies; 2+ messages in thread
From: Ben Woodard @ 2022-09-22  1:34 UTC (permalink / raw)
  To: libabigail

oops I also meant to say. The performance work had some modest benefit. 
The last time that I ran the test it was it took 23.5hrs on 18 cores. 
I'm not down to 14 cores and it took 22hrs.



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Where are we? (Sept 22 edition)
       [not found] ` <16c39076-2d42-0583-a013-7c9c0005f841@redhat.com>
@ 2022-09-25  7:05   ` Dodji Seketeli
  0 siblings, 0 replies; 2+ messages in thread
From: Dodji Seketeli @ 2022-09-25  7:05 UTC (permalink / raw)
  To: Ben Woodard via Libabigail; +Cc: Ben Woodard

Hello Ben,

Many thanks for running these very very important tests!

Ben Woodard via Libabigail <libabigail@sourceware.org> a écrit:

> Progress! With Dodji's latest patches I took the fails from the
> previous test and reran with the latest code. 80 packages which
> previously failed now pass.

Good to hear.

There are still issues, though, but I have a sense that we are in a much
better place than we were previously, when you first started to run
these monster tests.

> The class of "wrongly compared" seems to have continued to
> improve.

Cool. These were one of the area I focused the most on for this 2.1 release.

> There seem to be two classes of problems. One is just unspecified
> filtered symbols. The other class of problems are related to some kind
> of type confusion in C++.
>
> 1. crash
> 2. enlightenmentd
> 3. freefem++
> 4. gimp
> 5. harfbuzz
> 6. kdepim-addons
> 7. kf5-libktorrent
> 8. kf5-mailcommon
> 9. kscope
> 10. ldc
> 11. libcouchbase
> 12. nest
> 13. pocl
> 14. qt6-qtbase
> 15. smesh
> 16. tdlib
> 17. vsomeip3
>
> The asserts appear to be about the same but some of the line numbers
> have changed slightly but that is it. This affects 72 packages.

I have addressed some of the asserts in some latest patches that went in
today.

> The segv's are all gone yea! One interesting thing though is a few of
> the programs that were segv'ing are now returning 68. I have yet to 
> figure out what exit code 68 means:
>
> AusweisApp2 - Unknown 68
> kcov - Unknown 68
> bettercap - Unknown 68
> singularity - Unknown 68
> stargz-snapshotter - Unknown 68
> llvm11 - Unknown 68
> qt6-qtwayland - Unknown 68
> flang - Unknown 68

Several of these should now be gone, I think.

I'll go ahead and release 2.1.  I'll address the remaining issue later
unless someone beats me to it.

Thanks again.

[...]

Cheers,

-- 
		Dodji

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2022-09-25  7:05 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <22b479ce-09a5-52c6-24f9-d959310ce8f2@redhat.com>
2022-09-22  1:34 ` Where are we? Ben Woodard
     [not found] ` <16c39076-2d42-0583-a013-7c9c0005f841@redhat.com>
2022-09-25  7:05   ` Where are we? (Sept 22 edition) Dodji Seketeli

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).