[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: sbcl/2:2.2.3-1 [NMU] -- Common Lisp compiler and development system



Hi Kambiz,

Le jeudi 31 mars 2022 à 21:12 +0200, Kambiz Darabi a écrit :
> Dear all,
> 
> I am looking for a sponsor for package "sbcl":
> 
>  * Package name    : sbcl
>    Version         : 2:2.2.3-1
>  * URL             : http://www.sbcl.org
>  * License         : public-domain, Unicode, NTP~disclaimer, Apache-2.0, Expat, NTP, permissive-xerox, BSD-3-clause
>  * Vcs             : https://salsa.debian.org/common-lisp-team/sbcl
>    Section         : lisp

Unfortunately the autopkgtest fails, I had overlooked that problem in
the previous iteration of our discussion. See:
https://salsa.debian.org/common-lisp-team/sbcl/-/jobs/2624803

More precisely, the failing test is gc-cardmark.impure.lisp, and the
error message is:

// Running gc-cardmark.impure.lisp in COMPILE evaluator mode
::: Running :COMPACT-INSTANCE-HEADER-LAYOUT
/ cc  -fPIC -shared -I../src/runtime gc-testlib.c -o gc-testlib.so
gc-testlib.c:6:10: fatal error: sbcl.h: No such file or directory
    6 | #include "sbcl.h"
      |          ^~~~~~~~
compilation terminated.
::: UNEXPECTED-FAILURE :COMPACT-INSTANCE-HEADER-LAYOUT
    due to SB-INT:SIMPLE-FILE-ERROR:
        "Failed to find the TRUENAME of
/tmp/autopkgtest.XiRZ2K/build.4ud/src/tests/gc-testlib.so:

           No such file or directory"


This test succeeds at build time, so the problem comes from the
autopkgtest environment, which is slightly different. I had to adapt a
couple of tests so that they work under autopkgtests, see the various
patches under debian/patches/.

Note that if you can’t find a reasonable fix, a straightforward (though
not entirely satisfactory) solution would be to add this test to skip-
some-autopkgtests.patch.

Also note that running the testsuite in autopkgtests is useful even
though the testsuite is also run at build time. This can help detect
regressions caused by dependencies (since changes in dependencies don’t
trigger a rebuild).

P.S.: I remove the debian/2%2.2.3-1 tag in the git repository, since
that exact version is not going to be uploaded. Ideally, when you’re
requesting sponsorship, you should let the sponsor add the debian/*
tag.

P.S. #2: Don’t feel obliged to do a very formal RFS like you just did.
An informal request by email should be enough.

Thanks for your work,

-- 
⢀⣴⠾⠻⢶⣦⠀  Sébastien Villemot
⣾⠁⢠⠒⠀⣿⡁  Debian Developer
⢿⡄⠘⠷⠚⠋⠀  https://sebastien.villemot.name
⠈⠳⣄⠀⠀⠀⠀  https://www.debian.org

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: