Debian Bug report logs - #654381
icon: FTBFS on hurd-i386

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Svante Signell <svante.signell@telia.com>

To: Debian bug reporting <submit@bugs.debian.org>

Subject: icon: FTBFS on hurd-i386

Date: Tue, 03 Jan 2012 11:14:16 +0100

[Message part 1 (text/plain, inline)]

Source: icon
Version: 9.4.3-3
Severity: important
Tags: patch
User: debian-hurd@lists.debian.org
Usertags: hurd

Hi,

icon currently FTBFS on GNU/Hurd, see
https://buildd.debian.org/status/fetch.php?pkg=icon&arch=hurd-i386&ver=9.4.3-3&stamp=1306225911

The problems are a flaky fakeroot-tcp used by GNU/Hurd for the
Test-icont tests and a PATH setting issue in the Test-opts script.

The solution is to run the tests under the build target, not under the
binary target which is using fakeroot. The tests don't need to be
performed using fakeroot. The attached patch fixes these issues and
builds fine at least in the GNU/Hurd and the GNU/Linux i686+amd64
environments.

Thanks!

[fix_FTBFS4Hurd.diff (text/x-patch, attachment)]

Message #10 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Samuel Thibault <sthibault@debian.org>

To: svante.signell@telia.com, 654381@bugs.debian.org

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Tue, 3 Jan 2012 14:20:02 +0100

Svante Signell, le Tue 03 Jan 2012 11:14:16 +0100, a écrit :
> The problems are [...] a PATH setting issue in the Test-opts script.

What issue?

Thanks,
Samuel

Message #15 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Svante Signell <svante.signell@telia.com>

To: Samuel Thibault <sthibault@debian.org>

Cc: 654381@bugs.debian.org

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Tue, 03 Jan 2012 15:25:09 +0100

On Tue, 2012-01-03 at 14:20 +0100, Samuel Thibault wrote:
> Svante Signell, le Tue 03 Jan 2012 11:14:16 +0100, a écrit :
> > The problems are [...] a PATH setting issue in the Test-opts script.
> 
> What issue?

See the patch: Without adding . to the PATH the following happens:
error in startup code
/dev/fd/3: can't read interpreter file header

when execing the program hello, in tunix.c:  
ofile = "hello"; argv = "hello"; execv(ofile, argv).

Message #20 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Samuel Thibault <sthibault@debian.org>

To: Svante Signell <svante.signell@telia.com>

Cc: 654381@bugs.debian.org

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Tue, 3 Jan 2012 16:11:42 +0100

Svante Signell, le Tue 03 Jan 2012 15:25:09 +0100, a écrit :
> On Tue, 2012-01-03 at 14:20 +0100, Samuel Thibault wrote:
> > Svante Signell, le Tue 03 Jan 2012 11:14:16 +0100, a écrit :
> > > The problems are [...] a PATH setting issue in the Test-opts script.
> > 
> > What issue?
> 
> See the patch:

The patch only shows which source change you suggest, not what it is
supposed to fix.

> Without adding . to the PATH the following happens:
> error in startup code
> /dev/fd/3: can't read interpreter file header
> 
> when execing the program hello, in tunix.c:  
> ofile = "hello"; argv = "hello"; execv(ofile, argv).

I see. Problem is that it only cures the issue for the test. The -x
option is supposed to also work with the installed program, so I doubt
the maintainer will accept to cripple the test into accepting -x not
working without . in PATH.

As icont passes the right things to execv(), it rather looks like an
issue in execv itself.

Samuel

Message #25 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Svante Signell <svante.signell@telia.com>

To: Samuel Thibault <sthibault@debian.org>

Cc: 654381@bugs.debian.org

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Tue, 03 Jan 2012 16:31:19 +0100

On Tue, 2012-01-03 at 16:11 +0100, Samuel Thibault wrote:
> Svante Signell, le Tue 03 Jan 2012 15:25:09 +0100, a écrit :
> > On Tue, 2012-01-03 at 14:20 +0100, Samuel Thibault wrote:
> > > Svante Signell, le Tue 03 Jan 2012 11:14:16 +0100, a écrit :
> > > > The problems are [...] a PATH setting issue in the Test-opts script.
> 
> > Without adding . to the PATH the following happens:
> > error in startup code
> > /dev/fd/3: can't read interpreter file header
> > 
> > when execing the program hello, in tunix.c:  
> > ofile = "hello"; argv = "hello"; execv(ofile, argv).
> 
> I see. Problem is that it only cures the issue for the test. The -x
> option is supposed to also work with the installed program, so I doubt
> the maintainer will accept to cripple the test into accepting -x not
> working without . in PATH.
> 
> As icont passes the right things to execv(), it rather looks like an
> issue in execv itself.

OK, but the .-path problems would only be for GNU/Hurd, and since there
are no reports of execv problems for other architectures, then execv is
broken on Hurd, and not elsewhere?

Message #30 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Samuel Thibault <sthibault@debian.org>

To: Svante Signell <svante.signell@telia.com>

Cc: 654381@bugs.debian.org

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Tue, 3 Jan 2012 16:34:42 +0100

Svante Signell, le Tue 03 Jan 2012 16:31:19 +0100, a écrit :
> On Tue, 2012-01-03 at 16:11 +0100, Samuel Thibault wrote:
> > Svante Signell, le Tue 03 Jan 2012 15:25:09 +0100, a écrit :
> > > On Tue, 2012-01-03 at 14:20 +0100, Samuel Thibault wrote:
> > > > Svante Signell, le Tue 03 Jan 2012 11:14:16 +0100, a écrit :
> > > > > The problems are [...] a PATH setting issue in the Test-opts script.
> > 
> > > Without adding . to the PATH the following happens:
> > > error in startup code
> > > /dev/fd/3: can't read interpreter file header
> > > 
> > > when execing the program hello, in tunix.c:  
> > > ofile = "hello"; argv = "hello"; execv(ofile, argv).
> > 
> > I see. Problem is that it only cures the issue for the test. The -x
> > option is supposed to also work with the installed program, so I doubt
> > the maintainer will accept to cripple the test into accepting -x not
> > working without . in PATH.
> > 
> > As icont passes the right things to execv(), it rather looks like an
> > issue in execv itself.
> 
> OK, but the .-path problems would only be for GNU/Hurd, and since there
> are no reports of execv problems for other architectures, then execv is
> broken on Hurd, and not elsewhere?

Yes.

Samuel

Message #35 received at 654381@bugs.debian.org (full text, mbox, reply):

From: Samuel Thibault <sthibault@debian.org>

To: 654381@bugs.debian.org

Cc: Svante Signell <svante.signell@telia.com>

Subject: Re: Bug#654381: icon: FTBFS on hurd-i386

Date: Wed, 6 Jun 2012 23:31:22 +0200

[Message part 1 (text/plain, inline)]

Hello,

The exec part is now fixed in libc0.3 and hurd packages, so the PATH
part of the proposed patch is not needed any more.  Could you apply the
remainder (attached to this mail), which avoids an issue from fakeroot?

Samuel

[fix_FTBFS4Hurd.diff (text/plain, attachment)]

Message #42 received at 654381-close@bugs.debian.org (full text, mbox, reply):

From: Santiago Vila <sanvila@debian.org>

To: 654381-close@bugs.debian.org

Subject: Bug#654381: fixed in icon 9.4.3-5

Date: Tue, 09 Aug 2016 18:06:17 +0000

Source: icon
Source-Version: 9.4.3-5

We believe that the bug you reported is fixed in the latest version of
icon, which is due to be installed in the Debian FTP archive.

A summary of the changes between this version and the previous one is
attached.

Thank you for reporting the bug, which will now be closed.  If you
have further comments please address them to 654381@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.

Debian distribution maintenance software
pp.
Santiago Vila <sanvila@debian.org> (supplier of updated icon package)

(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@ftp-master.debian.org)


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Format: 1.8
Date: Tue, 09 Aug 2016 18:45:08 +0100
Source: icon
Binary: icont iconx iconc icon-ipl
Architecture: source
Version: 9.4.3-5
Distribution: unstable
Urgency: medium
Maintainer: Debian QA Group <packages@qa.debian.org>
Changed-By: Santiago Vila <sanvila@debian.org>
Description:
 icon-ipl   - Libraries for Icon, a high-level programming language
 iconc      - Compiler for Icon, a high-level programming language
 icont      - Interpreter for Icon, a high-level programming language
 iconx      - Executor for Icon, a high-level programming language
Closes: 654381 660466 821993 829133
Changes:
 icon (9.4.3-5) unstable; urgency=medium
 .
   * QA upload.
   * Package is orphaned (Bug #738875), set Maintainer to "Debian QA Group".
   * Do not run the tests as root. This used to make the build to fail on
     the Hurd. It does not anymore, but it's a bug anyway. Closes: #654381.
   * Ack NMUs. This should now build on sparc and mips. Closes: #660466.
   * Add build-arch and build-indep targets to debian/rules. Closes: #821993.
   * Apply patch from Reiner Herrmann to build reproducibly. Closes: #829133.
   * Switch to "3.0 (quilt)" source format.
   * General cleanup of debian/rules.
Checksums-Sha1:
 83f18bbc55156ad31a558f1e35abeb058d857a9b 1422 icon_9.4.3-5.dsc
 d657303ed38d2505b13ef6ca610d7a70fa413847 11164 icon_9.4.3-5.debian.tar.xz
Checksums-Sha256:
 0db4d1e1120ddca44d298c9c4408a1de0cdc22b5e6e019a22dcf369362f992b8 1422 icon_9.4.3-5.dsc
 32ad65a5d104177f9dc3ba7997e9292dc73597e22c33ccbb5259df515c23e577 11164 icon_9.4.3-5.debian.tar.xz
Files:
 41102d93d5cedb3885a7ee8acaa7c632 1422 devel optional icon_9.4.3-5.dsc
 7194b4033515810bfdc4b2f07bea1ce5 11164 devel optional icon_9.4.3-5.debian.tar.xz

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJXqhgxAAoJEEHOfwufG4sy0PUH/j2OCG+8crUrq/q3qpIDER8r
QNcCCv/0CwtDge4UjNKb6nSVR5PVORybUJGOBQWOqnvRTjScUOEOrSdE4++r6jqx
mIwrBpzmKxAQKZaFKqNVnMZWJ6vh5a0cIz/hmbCyUnF5jQkSVOsvqBd1q+459W/l
PJNbDexoIZ4as8ts00svNnXISd+B3bKs83Txfhad5RQRg4ji2+BZZevV4CNN++4T
eWYEySo79mwwWZieArhUzY08dItlr6NAvAWacZRd60NIdesmZrQLjnkwv+mvRO4O
pBN17e9RiZejckPZX2HM0fVFHS5LHGOV5gu3r2EjO7NnpfwObzlAk4i2tpbaEU0=
=9/9W
-----END PGP SIGNATURE-----

Send a report that this bug log contains spam.