Path: csiph.com!fu-berlin.de!bofh.it!news.nic.it!robomod From: Reinhard Tartler Newsgroups: linux.debian.bugs.rc Subject: Bug#1068755: docker.io: FTBFS: failing tests Date: Sun, 05 May 2024 21:10:02 +0200 Message-ID: References: X-Mailbox-Line: From debian-bugs-rc-request@lists.debian.org Sun May 5 19:09:13 2024 Old-Return-Path: X-Spam-Flag: NO X-Spam-Score: 1.653 Reply-To: Reinhard Tartler , 1068755@bugs.debian.org Resent-To: debian-bugs-dist@lists.debian.org Resent-Cc: Debian Go Packaging Team X-Debian-Pr-Message: followup 1068755 X-Debian-Pr-Package: src:docker.io X-Debian-Pr-Keywords: pending ftbfs X-Debian-Pr-Source: docker.io X-Gm-Message-State: AOJu0YwzvnUF+olvxTS7q2qn3hfZuj4ONV1/NHjXwf0fAX/3HRc7Cax3 EGydf4sR5kwkrNEBVsuoiqbZUWwwKJBjE2r/x0rXUoY46LDzy3VJHoSm5ZJEjRmAEe3U5j8yyRI 0+CB2u/TgeYslZ6sdsQDTQ1StNk/QtvVg X-Google-SMTP-Source: AGHT+IGPbqfOTdMqoX3wEWHTVoYmHeWnQr6RIWoabYlPXqu6QRk37JsTRIYNgC2cso0yaJ23c4mxisoWV/cI2BXda4g= X-Received: by 2002:a17:906:f84a:b0:a59:a01e:825f with SMTP id ks10-20020a170906f84a00b00a59a01e825fmr3697730ejb.29.1714935981414; Sun, 05 May 2024 12:06:21 -0700 (PDT) MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="000000000000fc233c0617b9a546" X-Debian-Message: from BTS X-Mailing-List: archive/latest/742235 List-ID: List-URL: Approved: robomod@news.nic.it Lines: 126 Organization: linux.* mail to news gateway Sender: robomod@news.nic.it X-Original-Cc: Shengjing Zhu X-Original-Date: Sun, 5 May 2024 15:06:09 -0400 X-Original-Message-ID: X-Original-References: <86b9e2dc-bac2-4a95-8d59-02556136c0b5@debian.org> Xref: csiph.com linux.debian.bugs.rc:363087 --000000000000fc233c0617b9a546 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Turns out that backporting https://github.com/moby/moby/commit/97921915a801dd82b1f5a70e0a69353539c1e3a= e.patch seems to actually make the test pass I'm not sure why that worked before, but I've pushed a backport of that patch to https://salsa.debian.org/go-team/packages/docker/-/commit/d33659365984dbb47= b6aac2836727e507c1ce737 to let the salsa pipeline work on it. I plan to make a team-upload tomorrow or later in the week. Please let me know if you have concerns or thoughts on this. Thanks -rt On Sun, May 5, 2024 at 11:20=E2=80=AFAM Reinhard Tartler wrote: > I've been looking at this test failure, but remain puzzled. > > Basically, the source for this test is here: > https://sources.debian.org/src/docker.io/20.10.25%2Bdfsg1-2/engine/distri= bution/xfer/download_test.go/#L364-L429. > This test is testing code in > https://sources.debian.org/src/docker.io/20.10.25+dfsg1-2/engine/distribu= tion/xfer/download.go > which has only two external dependencies > (golang-github-docker-distribution-dev) and > (golang-github-sirupsen-logrus-dev). > > I've checked the build log of the previous build at > https://buildd.debian.org/status/fetch.php?pkg=3Ddocker.io&arch=3Damd64&v= er=3D20.10.25%2Bdfsg1-2%2Bb3&stamp=3D1704672923&raw=3D0 > and note that the same test passes in that build. This also uses the same > package versions of golang-github-docker-distribution-dev=3D=3D2.8.2+ds1-= 1 and > golang-github-sirupsen-logrus-dev=3D=3D1.9.0-1 > > I also note that the old build was using golang 1.21, and the FTBFS was > introduced after upgrading unstable to golang 1.22. > > Shengjing, I believe you handled the recent update of the golang toolchai= n > from 1.21 to 1.22. Have you seen similar "mysterious" test failures that > resulted from using a newer golang compiler? > > I am considering just disabling this test, as the rest of the testsuite > seem to pass, but I'm getting nervous that this wouldn't just paper over = a > real issue. > > Thanks, > > > -- > regards, > Reinhard > --=20 regards, Reinhard --000000000000fc233c0617b9a546 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Turns out that backporting=C2= =A0https://github.com/moby/moby/commit/97921915a801dd82= b1f5a70e0a69353539c1e3ae.patch seems to actually make the test pass
=
I'm not sure why that worked before, but I've pushed a backport= of that patch to=C2=A0https://salsa.de= bian.org/go-team/packages/docker/-/commit/d33659365984dbb47b6aac2836727e507= c1ce737 to let the salsa pipeline work on it.

I plan to make a t= eam-upload tomorrow or later in=C2=A0the=C2=A0week. Please let me know if y= ou have concerns or thoughts on this.

Thanks
-rt

On Sun,= May 5, 2024 at 11:20=E2=80=AFAM Reinhard Tartler <siretart@gmail.com> wrote:
I've been looki= ng at this test failure, but remain puzzled.

Basically, the source f= or this test is here:=C2=A0https://sources.debian.org/src/docker.io/20.10.25%2Bdfs= g1-2/engine/distribution/xfer/download_test.go/#L364-L429. This test is= testing code in=C2=A0htt= ps://sources.debian.org/src/docker.io/20.10.25+dfsg1-2/engine/distribution/= xfer/download.go which has only two external dependencies (golang-githu= b-docker-distribution-dev) and (golang-github-sirupsen-logrus-dev).

= I've checked the build log of the previous build at=C2=A0https://buildd.debian.org/status/fetch.php?pkg=3Ddocker.io&= arch=3Damd64&ver=3D20.10.25%2Bdfsg1-2%2Bb3&stamp=3D1704672923&r= aw=3D0 and note that the same test passes in that build. This also uses= the same package versions of golang-github-docker-distribution-dev=3D=3D2.= 8.2+ds1-1 and golang-github-sirupsen-logrus-dev=3D=3D1.9.0-1

I also = note that the old build was using golang 1.21, and the FTBFS was introduced= after upgrading unstable to golang 1.22.

Shengjing, I believe you h= andled the recent update of the golang toolchain from 1.21 to 1.22. Have yo= u seen similar "mysterious" test failures that resulted from usin= g a newer golang compiler?

I am considering just disabling this test= , as the rest of the testsuite seem to pass, but I'm getting nervous th= at this wouldn't just paper over a real issue.

Thanks,


--
regards,
=C2=A0 =C2=A0 Reinhard<= /div>


--
re= gards,
=C2=A0 =C2=A0 Reinhard
--000000000000fc233c0617b9a546--