Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.maint.python > #14510 > unrolled thread

review for python-ssdpy

Started byJeroen Ploemen <jcfp@debian.org>
First post2022-10-03 14:20 +0200
Last post2022-10-04 18:50 +0200
Articles 4 — 2 participants

Back to article view | Back to linux.debian.maint.python


Contents

  review for python-ssdpy Jeroen Ploemen <jcfp@debian.org> - 2022-10-03 14:20 +0200
    Re: review for python-ssdpy Ben Westover <kwestover.kw@gmail.com> - 2022-10-03 18:00 +0200
      Re: review for python-ssdpy Jeroen Ploemen <jcfp@debian.org> - 2022-10-04 15:00 +0200
        Re: review for python-ssdpy Ben Westover <kwestover.kw@gmail.com> - 2022-10-04 18:50 +0200

#14510 — review for python-ssdpy

FromJeroen Ploemen <jcfp@debian.org>
Date2022-10-03 14:20 +0200
Subjectreview for python-ssdpy
Message-ID<FcsAp-bLQN-7@gated-at.bofh.it>

[Multipart message — attachments visible in raw view] — view raw

Hi Ben,

* control: missing build-dep on python3-all? Needed to run the
  testsuite against all supported Python version rather than just the
  default one.

* watch: regex "(\d\.\d.\d)" is overly specific, could easily miss
  upstream versions such as a hypothetical "1.5" or "1.4.10".

Package ftbfs in pbuilder with errors in the testsuite; many of the
failing tests are marked by upstream as incompatible with github ci,
probably for similar reasons they fail in pbuilder, particularly
assumptions regarding the availability and configuration of the
network.

See log at https://paste.debian.net/hidden/daadfbea/

[toc] | [next] | [standalone]


#14511

FromBen Westover <kwestover.kw@gmail.com>
Date2022-10-03 18:00 +0200
Message-ID<Fcw1k-bNH3-3@gated-at.bofh.it>
In reply to#14510

[Multipart message — attachments visible in raw view] — view raw

Hello,

On 10/3/22 08:12, Jeroen Ploemen wrote:
> * control: missing build-dep on python3-all? Needed to run the
>    testsuite against all supported Python version rather than just the
>    default one.

Thanks, added.

> * watch: regex "(\d\.\d.\d)" is overly specific, could easily miss
>    upstream versions such as a hypothetical "1.5" or "1.4.10".

Fixed.

> Package ftbfs in pbuilder with errors in the testsuite; many of the
> failing tests are marked by upstream as incompatible with github ci,
> probably for similar reasons they fail in pbuilder, particularly
> assumptions regarding the availability and configuration of the
> network.
> 
> See log at https://paste.debian.net/hidden/daadfbea/

Thanks, I didn't catch that since I use sbuild instead of pbuilder. I 
have set the environment variable CI = true which should disable those 
tests. This and above fixes are pushed to Salsa.

--
Ben Westover

[toc] | [prev] | [next] | [standalone]


#14513

FromJeroen Ploemen <jcfp@debian.org>
Date2022-10-04 15:00 +0200
Message-ID<FcPGF-bZyc-19@gated-at.bofh.it>
In reply to#14511

[Multipart message — attachments visible in raw view] — view raw

On Mon, 3 Oct 2022 11:56:30 -0400
Ben Westover <kwestover.kw@gmail.com> wrote:

> Thanks, I didn't catch that since I use sbuild instead of pbuilder.
> I have set the environment variable CI = true which should disable
> those tests.

Unfortunately, it doesn't.

The change suggested below does make 'CI=true' take effect, but that
still leaves a few failing tests not marked by upstream. Probably
easiest to exclude those remaining troublemakers by instructing
pybuild to set a pytest option, e.g. PYBUILD_TEST_ARGS = -k 'not ...'.

--- a/debian/rules
+++ b/debian/rules
@@ -1,7 +1,6 @@
 #! /usr/bin/make -f
 
 export PYBUILD_NAME = ssdpy
-export CI = "true" # Skip tests which require network
 
 %:
        dh $@ --with python3,sphinxdoc --buildsystem=pybuild
@@ -11,3 +10,7 @@ execute_after_dh_auto_build:
 
 execute_after_dh_auto_clean:
        rm -rf debian/html
+
+override_dh_auto_test:
+       # Skip tests which require network
+       CI=true dh_auto_test

[toc] | [prev] | [next] | [standalone]


#14514

FromBen Westover <kwestover.kw@gmail.com>
Date2022-10-04 18:50 +0200
Message-ID<FcThf-c1La-1@gated-at.bofh.it>
In reply to#14513

[Multipart message — attachments visible in raw view] — view raw

On 10/4/22 8:50 AM, Jeroen Ploemen wrote:
>> Thanks, I didn't catch that since I use sbuild instead of pbuilder.
>> I have set the environment variable CI = true which should disable
>> those tests.
> 
> Unfortunately, it doesn't.
> 
> The change suggested below does make 'CI=true' take effect, but that
> still leaves a few failing tests not marked by upstream. Probably
> easiest to exclude those remaining troublemakers by instructing
> pybuild to set a pytest option, e.g. PYBUILD_TEST_ARGS = -k 'not ...'.

I was going to, but it looks like practically all of the tests require 
network in some form, so I've just disabled tests altogether.

Thanks,
--
Ben Westover

[toc] | [prev] | [standalone]


Back to top | Article view | linux.debian.maint.python


csiph-web