[llvm-dev] dexter failures in cross-project-tests

via llvm-dev llvm-dev at lists.llvm.org
Thu Oct 28 10:14:48 PDT 2021

Further research shows that no bot uses “all” targets; the bot clang-x64-windows-msvc is listed as running debuginfo-tests, but that ninja target became invalid after the move to cross-project-tests.  It’s the only bot I see in the zorg config file that has that target.  So, apparently the missing piece of work is to get that bot switched over to the new target (and make sure it’s still passing).
@rnk is listed as the owner of that bot.  With some notice, we (Sony) should be able to provide assistance if the switch doesn’t Just Work.

I want to apologize for all the fuss and bother here.  We do want things to be working.  Is it worth filing a PR for this?  We have an internal ticket already to get this taken care of.

From: llvm-dev <llvm-dev-bounces at lists.llvm.org> On Behalf Of via llvm-dev
Sent: Thursday, October 28, 2021 9:18 AM
To: dblaikie at gmail.com; Cazalet-Hyams, Orlando <orlando.hyams at sony.com>; llvm-dev at lists.llvm.org
Cc: Weaver, Tom <Tom.Weaver at sony.com>
Subject: Re: [llvm-dev] dexter failures in cross-project-tests

> without any buildbots running them

Perhaps naively, I thought cross-project-tests (which includes dexter) was part of check-all, and so would be running on any bots that used check-all.  Do bots not use check-all?  Internally it looks like our Linux build does exactly that, and I see dexter tests passing there.

The goal of providing dexter to the community was indeed to make it useful to others, which obviously includes making sure it works.  Your concerns are completely fair and we are actively trying to figure out what’s going on.

From: llvm-dev <llvm-dev-bounces at lists.llvm.org<mailto:llvm-dev-bounces at lists.llvm.org>> On Behalf Of David Blaikie via llvm-dev
Sent: Wednesday, October 27, 2021 4:57 PM
To: Cazalet-Hyams, Orlando <orlando.hyams at sony.com<mailto:orlando.hyams at sony.com>>
Cc: llvm-dev <llvm-dev at lists.llvm.org<mailto:llvm-dev at lists.llvm.org>>; Weaver, Tom <Tom.Weaver at sony.com<mailto:Tom.Weaver at sony.com>>
Subject: Re: [llvm-dev] dexter failures in cross-project-tests

On Wed, Oct 27, 2021 at 7:13 AM <orlando.hyams at sony.com<mailto:orlando.hyams at sony.com>> wrote:
Hi David,

> I'm trying to add some test coverage to cross-project-tests (related to the simplified template names work), and figured I'd start by making sure everything passes. I fixed some pretty printer issues (a36032345ed823414fcee0f9aed43628f4320001 - though that makes me wonder: is any buildbot running these tests, because they look like they've been failing for a while?) and then some dexter issues.

Apple's lldb bot (https://green.lab.llvm.org/green/view/LLDB/job/lldb-cmake<https://urldefense.com/v3/__https:/green.lab.llvm.org/green/view/LLDB/job/lldb-cmake__;!!JmoZiZGBv3RvKRSx!u1SYKf5CJom-1aN_QxnGYzdOu9NswlXNOxtGuYPL6blgRTBBow5XNqlH6PQZEQ5yRg$>) runs the cross-project-tests. Looking at it now though, it looks like the job status isn't affected by these tests if they fail, a bunch of the dexter feature_tests appear to contain "UNSUPPORTED: system-darwin", plus - picking a "passing" build (https://green.lab.llvm.org/green/view/LLDB/job/lldb-cmake/37155<https://urldefense.com/v3/__https:/green.lab.llvm.org/green/view/LLDB/job/lldb-cmake/37155__;!!JmoZiZGBv3RvKRSx!u1SYKf5CJom-1aN_QxnGYzdOu9NswlXNOxtGuYPL6blgRTBBow5XNqlH6PTLY3mMQQ$>) - a load of dexter tests fail due to environment issues. All in all I'd say that doesn't really count, and I'm not aware of any other bots running these tests.

Given Sony contributed dexter and these tests - do you folks have any interest/resources invested in keeping these tests green? If not, maybe they should be removed? You can keep them to run when you want, but without any buildbots running them they don't seem sufficiently valued?

> Anyone know what these dexter failures are about? The first one doesn't even seem to have the command line arguments right, which seems concerning/like something's quite broken. The second one gives not much meaningful/actionable output - and I tried running dexter directly with --verbose and seeing what it was doing (it didn't print much, but said it couldn't print the value of a local variable when it was passed to printf?), tried building the binary with the cflags that are passed to dexter and the binary ran correctly/was debuggable/I could print the value of the variable - seems like dexter could be more informative both while running under lit/the default configuration, and otherwise?

Looks like the first test fails because of changes in D109833 - I've put a fix up here D112624.


 I'm not sure about the second failure without digging in deeper. One of us (cc Tom, Jeremy) can possibly take a look on or after Friday - we could XFAIL it until we get the chance?

Sure, whatever works for you - seems important to not have standing failures like this one way or another.

- Dave

N.B. Stephen is away at the moment.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20211028/a9138fce/attachment.html>

More information about the llvm-dev mailing list