Processed: src:sbcl: fails to migrate to testing for too long

[Debian Bug Tracking System]

close -1 2:2.4.9+git20241010.1.465757f12-1

Bug #1086128 [src:sbcl] src:sbcl: fails to migrate to testing for too long
Marked as fixed in versions sbcl/2:2.4.9+git20241010.1.465757f12-1.
Bug #1086128 [src:sbcl] src:sbcl: fails to migrate to testing for too long
Marked Bug as done

[Douglas Katzman]

no size function for object at 0x100009bf80 (widetag 0x33)
(in particular, no GC debugging output)
Does it give the opportunity to poke around in 'ldb' ? It's got to be

doing something before exiting

[Stas Boukarev]

I have random corruptions when GCing.
./run-tests.sh compiler-2.pure.lisp --gc-stress
::: Running (DEBUG :UNUSED-TN-VERY-LONG-ARGLIST)
./subr.sh: line 6: 94966 Segmentation fault "$SBCL_RUNTIME"
--lose-on-corruption --core "$SBCL_CORE" $SBCL_ARGS --eval "(setf
sb-ext:*evaluator-mode* :${TEST_SBCL_EVALUATOR_MODE:-compile})" "$@"

Sometimes it's sigill.

On Wed, Nov 20, 2024 at 7:28 PM Douglas Katzman via Sbcl-devel

no size function for object at 0x100009bf80 (widetag 0x33)
(in particular, no GC debugging output)

Does it give the opportunity to poke around in 'ldb' ? It's got to be doing something before exiting
_______________________________________________
Sbcl-devel mailing list
https://lists.sourceforge.net/lists/listinfo/sbcl-devel

[Sean Whitton]

Hello,

no size function for object at 0x100009bf80 (widetag 0x33)
(in particular, no GC debugging output)
Does it give the opportunity to poke around in 'ldb' ? It's got to be doing something
before exiting

The test execution environment is non-interactive, so we drive it using
'sbcl --script'. Is there some way I could feed in some ldb commands in
advance, and are there some you would suggest?

[Sean Whitton]

Hello Paul,

I *think* you're aware, but maybe I'm mixing up history: while this is true,
the operators of ci.d.n (me and terceiro) can give access to a testbed on the
actual hardware where we run our tests. We'd need to agree on a date/time
because in the way it works, either of us needs to remain on-line at that
time. Normally we'd give you the testbed where the test already ran, but
failed, for immediate inspection or other use you find necessary for
debugging.

I didn't know about this, so thank you for mentioning it.

I don't think that I am willing to schedule volunteer time like this for
this particular bug. I have an RFH out for sbcl, and I'm selective in
what I work on.

[Sean Whitton]

Hello Release Team,

I don't think I can do anything about this bug. Upstream aren't able to
reproduce it. I don't think it makes sense to block sbcl's migration on
this. I think that either we should ignore the cl-ironclad test failure
on ppc64el or remove sbcl on that architecture, if you think that would
be warranted.

Please let me know if you would be okay with overriding the test
failure.

Thanks.

[Sean Whitton]

Hello Release Team,

I don't think I can do anything about this bug. Upstream aren't able to
reproduce it. I don't think it makes sense to block sbcl's migration on
this. I think that either we should ignore the cl-ironclad test failure
on ppc64el or remove sbcl on that architecture, if you think that would
be warranted.

Please let me know if you would be okay with overriding the test
failure.

Thanks.

[Sean Whitton]

Hello,

Hi ppc64el porters, Sean,
@porters, can you please have a look at the ppc64el autopkgtest regressions
caused by sbcl [1]?

Post by Sean Whitton
I don't think I can do anything about this bug. Upstream aren't able to
reproduce it. I don't think it makes sense to block sbcl's migration on
this. I think that either we should ignore the cl-ironclad test failure
on ppc64el or remove sbcl on that architecture, if you think that would
be warranted.

no size function for object at 0x100009bf80 (widetag 0x33)

Right -- I think the error occurs when sbcl tries to load cl-ironclad,
and cl-postmodern depends on cl-ironclad, so it's essentialy the same
error.

(Though, typing this, I realise that I may have assumed this without
adequate evidence.)

Post by Sean Whitton
Please let me know if you would be okay with overriding the test
failure.

I assuming you didn't contact the ppc64el porters yet (apologies if you
did). Shall we give them a chance to investigate?

I didn't. Many thanks for doing so -- I'm happy to wait.

[Paul Gevers]

Hi,

How severe do you think this error is? Does it indicate that sbcl is
useless on ppc64el or is this a niche use case?

I would doubt that it means that sbcl is useless on ppc64el because of
how upstream can't reproduce the problem.

Ack. Let's give the porters a few more days, and otherwise accept the
regression of cl-ironclad and cl-postmodern on ppc64el.

My offer of access to a testbed on the ci.d.n infrastucture still stands.

Paul