[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

One last OPEN QUESTION (was: [discuss] Required Python Versions)

From: Greg Stein <gstein_at_gmail.com>
Date: Thu, 9 Apr 2009 17:28:35 +0200

See end of this email for the remaining open question.

2009/4/9 Arfrever Frehtes Taifersar Arahesis <arfrever.fta_at_gmail.com>:
> 2009-04-09 11:35:38 Greg Stein napisał(a):
>> On Thu, Apr 9, 2009 at 10:54, Arfrever Frehtes Taifersar Arahesis
>> <arfrever.fta_at_gmail.com> wrote:
>> > 2009-04-07 19:14:57 Greg Stein napisał(a):
>> >> Everybody:
>> >>
>> >> The current trend of this thread is to follow the plan below.
>
> IMHO this thread was about which Python versions are supported, not
> about what to do with code related to currrently unsupported versions.

I thought it was pretty obvious that it also meant "get rid of the
stuff which is creating an unnecessary maintenance cost", but I can
see where that may not have been clear.

>...
>> I wasted time tracking down a serf test failure a couple days ago,
>> only to find it was because the output had been converted to Unicode,
>> but the output verification pattern was non-Unicode. Thus, it couldn't
>> find it.
>
> I can't reproduce it:

Maybe because of your change in r37128.

Doesn't matter whether it works *now* ... it failed before, and it
should not have. This useless support for 3.0 is creating problems.
That is a proven statement based on my experience from a couple days
ago.

What is the next problem that I'm going to run into? The 3.0 support
has destabilized the test suite.

>...
>> r37128 brings in even more if/else branches for 3.0, so it is *still*
>> adding complexity to a codebase which doesn't need it.
>
> The minimal amount of additional complexity exists in very rarely changed
> code and doesn't cause any problems. There are no technical reasons to
> remove it.

Very rarely changed? I was just in there last month working on wc.py,
adding sandbox.py, redoing the testcase logic, whacking tree.py ...
etc. I was touching *everything*. And I plan on continuing that work
next time I'm feeling up for a break from wc-ng.

If I have to deal with 3.0 code, then my work is more difficult. There
is no benefit to keeping it, but there is an obvious cost.

>> It is all coming out later today.
>
> Please post patches for discussion before removing any code.

No.

This thread is the discussion. So far, the community has said "we do
not want to support 3.0 in the test suite, or other scripts for the
svn developers." In fact, except for yourself, that position has been
unanimous.

That said, I can see where it may not have been obvious that I was
going to *remove* 3.0 support from our codebase. Therefore, I'm going
to leave this open until Sunday (based on my own availability, may as
well give it until then). So. Let's make this clear:

DEVELOPERS: is anybody opposed to my *removal* of 3.0 support from the
dev Python scripts?

Cheers,
-g

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1615629
Received on 2009-04-09 21:35:12 CEST

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.