[Tracker-discuss] [issue270] Upgrade to 1.4.8

[Daniel Diniz]

Daniel Diniz <ajaksu at gmail.com> added the comment:

Martin v. Löwis wrote:
>> Should I keep this change (lowercasing existing POSTs) out of the 1.4.8 patch?
>
> Yes, please. A "it now validates" patch would be better.

OK, here's an updated, cleaner patch.

I do have an incomplete "it now validates" local branch, but it always seems to
be better to wait until most template changes land before working on it. I hope
I can include it on a new work plan so I don't block myself with incompatible
patches.

> Ok, patching detectors to match more closely the upstream reactors
> is good (i.e. if a fragment is in 1.4.8, it has a place in this patch).
>
> Stuff that is *not* in the upstream code does not belong into this
> patch. 

OK, new template patch uses this rule, no big changes. I haven't included id
strings and copyright notices from upstream, are they desirable?

The roundup patch has a couple of trunk changes that aren't in 1.4.8, should I
split it in 1.4.8 and 1.4.9a? 

> Wrt. 2.3 support, I think any effort in providing it is wasted.
>
>> IMHO, both explicitly targeting a specific (set of) version(s) or
>> using the anypy helper could help us make our detectors better and
>> less prone to bit-rot. So, what would you prefer: target 2.4 and above
>> (2.x), target  2.4-2.6, rely on anypy, something else?
>
> I disagree that the python-dev tracker is the place to start
> a generic reusable template.

No problem, let's forget the 'more generic template' idea. But having a set of
target versions would make my work easier. A lower bound (2.4?) makes it easier
to decide which features can be used, but an upper bound (2.6? 2.x for x >= 7?)
also helps in coding future-ready code.

> Feel free to contribute it to roundup,
> and perhaps talk roundup maintainers into making it a new template,
> but leave this project please out of our instances.

Stefan Seefeld has already started this, and IIUC has added many desirable
features to his python-dev-based template, like support for tasks and
milestones. When it lands upstream I'll see what we can reuse.

I'm submitting our local roundup-src changes upstream, so they can decide to
integrate them or not. Sometimes (like with the unlink auditor), I'm also
submitting parallel patches for the classic template. 

> The tracker instance is *not* reusable. It is tightly integrated
> with the rest of python.org in terms of layout, and it should
> stay that way.

OK. I'll keep a set of local patches to use in http://bot.bio.br/python-dev (a
test instance) and in the experimental instance when it's live.

> BTW, parallel contributions to the Jython tracker would be appreciated (although
> we could also defer such tasks to the Jython guys).
>
> Likewise, the meta and setuptools instances might need some care; this should be
> easy since they really use the original templates very much.

OK, I'll take a look at each instance to see what and how I contribute to them.
Are we free to update their templates and detectors or should I also ping
someone else?

Expect a tracker-discuss thread with a draft updated roadmap in the week after
GSoC students are announced. That way I can coordinate with any overlapping
projects to avoid conflicts.

_______________________________________________________
PSF Meta Tracker <metatracker at psf.upfronthosting.co.za>
<http://psf.upfronthosting.co.za/roundup/meta/issue270>
_______________________________________________________
-------------- next part --------------
A non-text attachment was scrubbed...
Name: python-dev_1.4.8_2.diff
Type: text/x-diff
Size: 4367 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/tracker-discuss/attachments/20090413/e8237898/attachment.diff>