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


Groups > linux.debian.maint.python > #9622 > unrolled thread

removing python3-trollius

Started byMichael Hudson-Doyle <michael.hudson@canonical.com>
First post2017-07-10 11:50 +0200
Last post2017-07-13 05:30 +0200
Articles 4 — 2 participants

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


Contents

  removing python3-trollius Michael Hudson-Doyle <michael.hudson@canonical.com> - 2017-07-10 11:50 +0200
    Re: removing python3-trollius Tristan Seligmann <mithrandi@mithrandi.net> - 2017-07-10 12:10 +0200
      Re: removing python3-trollius Michael Hudson-Doyle <michael.hudson@canonical.com> - 2017-07-13 02:10 +0200
        Re: removing python3-trollius Michael Hudson-Doyle <michael.hudson@canonical.com> - 2017-07-13 05:30 +0200

#9622 — removing python3-trollius

FromMichael Hudson-Doyle <michael.hudson@canonical.com>
Date2017-07-10 11:50 +0200
Subjectremoving python3-trollius
Message-ID<u1Duh-3kz-3@gated-at.bofh.it>

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

Hi,

python3-trollius fails tests with Python 3.6 and rather than think about
why, I think it makes more sense to stop building trollius for Python 3.
It's reason for existence is to backport asyncio to older Python 3 versions
(and Python 2, but I'm not talking about the python 2 package here). As sid
only has versions of Python 3 that very much do have asyncio, it's a bit
pointless now, and in fact I think all reverse-dependencies bar one
actually conditionally only import trollius if asyncio is not found and so
the dependencies can simply be dropped. The one exception is trollius-redis
which has no reverse dependencies and seems to be dead upstream and so
could be removed (or converted to build for Python 2 only).

Thoughts?

Cheers,
mwh

[toc] | [next] | [standalone]


#9623

FromTristan Seligmann <mithrandi@mithrandi.net>
Date2017-07-10 12:10 +0200
Message-ID<u1DNF-3IW-17@gated-at.bofh.it>
In reply to#9622

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

On Mon, 10 Jul 2017 at 11:45 Michael Hudson-Doyle <
michael.hudson@canonical.com> wrote:

>
> python3-trollius fails tests with Python 3.6 and rather than think about
> why, I think it makes more sense to stop building trollius for Python 3.
>

Sounds reasonable to me!

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


#9624

FromMichael Hudson-Doyle <michael.hudson@canonical.com>
Date2017-07-13 02:10 +0200
Message-ID<u2zRD-6HJ-3@gated-at.bofh.it>
In reply to#9623

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

On 10 July 2017 at 21:52, Tristan Seligmann <mithrandi@mithrandi.net> wrote:

> On Mon, 10 Jul 2017 at 11:45 Michael Hudson-Doyle <
> michael.hudson@canonical.com> wrote:
>
>>
>> python3-trollius fails tests with Python 3.6 and rather than think about
>> why, I think it makes more sense to stop building trollius for Python 3.
>>
>
> Sounds reasonable to me!
>

So a wrinkle here is that trollius-redis was packaged as it is a dependency
of pagure. Pagure is currently Python 2 only so removing python3-trollius
and python3-trollius-redis will not impede that effort in the short term,
but it might make life harder as and when pagure does switch to Python 3.
Of course, we could make the (reasonable sounding) bet that if pagure
switches to Python 3, it will also switch to asyncio rather than depending
on the deprecated trollius.

What do you think? Having written the above out, I think it still makes
sense to proceed with my plan. I'll start filing bugs and patches and
looking for sponsors...

Cheers,
mwh

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


#9625

FromMichael Hudson-Doyle <michael.hudson@canonical.com>
Date2017-07-13 05:30 +0200
Message-ID<u2CZb-cr-15@gated-at.bofh.it>
In reply to#9624

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

On 13 July 2017 at 11:46, Michael Hudson-Doyle <michael.hudson@canonical.com
> wrote:

> On 10 July 2017 at 21:52, Tristan Seligmann <mithrandi@mithrandi.net>
> wrote:
>
>> On Mon, 10 Jul 2017 at 11:45 Michael Hudson-Doyle <
>> michael.hudson@canonical.com> wrote:
>>
>>>
>>> python3-trollius fails tests with Python 3.6 and rather than think about
>>> why, I think it makes more sense to stop building trollius for Python 3.
>>>
>>
>> Sounds reasonable to me!
>>
>
> So a wrinkle here is that trollius-redis was packaged as it is a
> dependency of pagure. Pagure is currently Python 2 only so removing
> python3-trollius and python3-trollius-redis will not impede that effort in
> the short term, but it might make life harder as and when pagure does
> switch to Python 3. Of course, we could make the (reasonable sounding) bet
> that if pagure switches to Python 3, it will also switch to asyncio rather
> than depending on the deprecated trollius.
>
> What do you think? Having written the above out, I think it still makes
> sense to proceed with my plan. I'll start filing bugs and patches and
> looking for sponsors...
>

The bugs and patches are all at

https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=no-more-python3-trollius;users=debian-python@lists.debian.org

if someone wants to get their sponsoring boots on :)

Cheers,
mwh

[toc] | [prev] | [standalone]


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


csiph-web