This is a reply to a message from last May that I'd set aside at the time
because I wanted to think about it, and then it got buried in my to-do
list for all this time. Sorry about that. I've been thinking about it
off and on since then, at least.
Stefano Zacchiroli z...@debian.org writes:
On Thu, May 14, 2009 at 10:12:56PM +1100, Romain Beauxis wrote:
Le Thursday 14 May 2009 06:47:48 Sylvain Le Gall, vous avez écrit :
This kind of external plugin architecture can be a very good way for
packaging team to enhance their policy and their QA.
Just to add my 2 cents: If the
Le Thursday 14 May 2009 22:44:24 Stefano Zacchiroli, vous avez écrit :
Just to add my 2 cents: If the policy is one package + one lintian
version = deterministic checks, then why not adding a flag in the
package's control file that specifies which additional set of tests
should be run when
On Thu, May 14, 2009 at 11:01:53PM +1100, Romain Beauxis wrote:
urrrmmm, not really. The point here is also the extra need of
some language-specific tool which do not appear as lintian
dependency. So, even if I've a sympathy for your proposal,
reproducibility will not be ensured
On Tue, May 12, 2009 at 05:46:18PM +0100, Adam D. Barratt wrote:
I'm not Russ, obviously, but fwiw with my Lintian co-maintainer hat
on...
Thanks to both Adam and Russ for the clear answers.
My initial thought is that doing so would break reproducibility of
Lintian results. Maintaining that
On 13-05-2009, Stefano Zacchiroli z...@debian.org wrote:
On Tue, May 12, 2009 at 11:08:37AM -0700, Russ Allbery wrote:
However I see a difference on scope. We, Debian OCaml Maintainers,
want a tool to run our own tests, use it for our own team-specific QA,
and give it to our packagers as a
Stefano Zacchiroli a écrit :
This check is quite easy using the ocamlobjinfo tool: it prints Force
custom: YES when given a faulty .cma.
... but I doubt that we can rely on external packages from lintian
checks (lintian maintainers: can we?). So I suggest implementing the
test in pure Perl,
Zack wrote:
[ replying to Russ' post, which contains the interesting part to
support Stephane's reply ]
I'm not Russ, obviously, but fwiw with my Lintian co-maintainer hat on...
In this particular case it will be /usr/bin/ocamlobjinfo. The lintian
test will do nothing, silently, if the
Package: lintian
Severity: wishlist
The OCaml compiler enables linking code in various ways: most notably
bytecode and nativecode. The bytecode executables though can be linked
together with the interpreter (the so called custom mode), generating
ELF executables which can not be stripped. Since
Stefano Zacchiroli wrote:
As such, we would like to add a lintian check to warn against OCaml
custom mode executable. They can easily detected by looking for a magic
number at the end of the file, as described in the forwarded mail from
upstream.
Libraries can force custom mode executables.
On Sun, Sep 07, 2008 at 04:51:07PM +0200, Stéphane Glondu wrote:
This check is quite easy using the ocamlobjinfo tool: it prints Force
custom: YES when given a faulty .cma.
... but I doubt that we can rely on external packages from lintian
checks (lintian maintainers: can we?). So I suggest
Stefano Zacchiroli [EMAIL PROTECTED] writes:
On Sun, Sep 07, 2008 at 04:51:07PM +0200, Stéphane Glondu wrote:
This check is quite easy using the ocamlobjinfo tool: it prints Force
custom: YES when given a faulty .cma.
... but I doubt that we can rely on external packages from lintian
checks
12 matches
Mail list logo