Add a warning if the current version of python is not 2.5

Add a warning if the current version of python is not 2.5

R=dpranke@chromium.org
BUG=
TEST=

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=80212

[Dirk Pranke, 2011-04-01 18:53:56]

I don't like this patch. I understand what you're trying to accomplish, but it's
at least as important that depot_tools be tested on 2.6. I think it's better to
run the tests on both versions if you can, or set up automated bots to test the
changes on the different configurations instead.

[M-A Ruel, 2011-04-01 19:32:40]

On 2011/04/01 18:53:56, Dirk Pranke wrote:
> I don't like this patch. I understand what you're trying to accomplish, but
it's
> at least as important that depot_tools be tested on 2.6. I think it's better
to
> run the tests on both versions if you can, or set up automated bots to test
the
> changes on the different configurations instead.

In general, python 2.6 won't break stuff that works on 2.5. The reverse is not
true.

Also, this is just a message, not even a warning.

I've fixed all the scripts to pass cleanly on pylint with python 2.5 and use
/usr/bin/env python everywhere so it's symlink friendly, without needing to run
"hash -r".

[Dirk Pranke, 2011-04-01 19:45:14]

There are two other, more important reasons for me not wanting to encourage
people to run Python 2.5 by default.

1) new-run-webkit-tests can use the multiprocessing module in Python 2.6, which
speeds up the tests enormously and makes things more stable.

2) Because of (1) I actually want to bundle Python 2.6 into third_party
somewhere so that we can run it by default on Leopard. It's already what we have
on SL, Lucid, and all of the Win bots, and Hardy is about to go away. 

So this feels like a step in the wrong direction.

That said, maybe there are other repos and bots that will need to support 2.5
for a longer period of time?

The env changes are fine, however.

[M-A Ruel, 2011-04-01 19:51:49]

I don't want to encourage anyone to run python 2.5 by default. I just want to
make sure that anyone hacking *on depot_tools* should test it with python 2.5.

That's different.

M-A

On 2011/04/01 19:45:14, Dirk Pranke wrote:
> There are two other, more important reasons for me not wanting to encourage
> people to run Python 2.5 by default.
> 
> 1) new-run-webkit-tests can use the multiprocessing module in Python 2.6,
which
> speeds up the tests enormously and makes things more stable.
> 
> 2) Because of (1) I actually want to bundle Python 2.6 into third_party
> somewhere so that we can run it by default on Leopard. It's already what we
have
> on SL, Lucid, and all of the Win bots, and Hardy is about to go away. 
> 
> So this feels like a step in the wrong direction.
> 
> That said, maybe there are other repos and bots that will need to support 2.5
> for a longer period of time?
> 
> The env changes are fine, however.

[Dirk Pranke, 2011-04-01 20:05:21]

Ok, fair enough.

LGTM, although I would still prefer that this was done
programmatically rather than by user nudging.

-- Dirk

On Fri, Apr 1, 2011 at 12:51 PM,  <maruel@chromium.org> wrote:
> I don't want to encourage anyone to run python 2.5 by default. I just want
> to
> make sure that anyone hacking *on depot_tools* should test it with python
> 2.5.
>
> That's different.
>
> M-A
>
> On 2011/04/01 19:45:14, Dirk Pranke wrote:
>>
>> There are two other, more important reasons for me not wanting to
>> encourage
>> people to run Python 2.5 by default.
>
>> 1) new-run-webkit-tests can use the multiprocessing module in Python 2.6,
>
> which
>>
>> speeds up the tests enormously and makes things more stable.
>
>> 2) Because of (1) I actually want to bundle Python 2.6 into third_party
>> somewhere so that we can run it by default on Leopard. It's already what
>> we
>
> have
>>
>> on SL, Lucid, and all of the Win bots, and Hardy is about to go away.
>
>> So this feels like a step in the wrong direction.
>
>> That said, maybe there are other repos and bots that will need to support
>> 2.5
>> for a longer period of time?
>
>> The env changes are fine, however.
>
>
>
> http://codereview.chromium.org/6791018/
>

[M-A Ruel, 2011-04-01 20:22:02]

On 2011/04/01 20:05:21, Dirk Pranke wrote:
> Ok, fair enough.
> 
> LGTM, although I would still prefer that this was done
> programmatically rather than by user nudging.

I agree but the presubmit script cannot install anything so that's really just a
best effort here. I thought about finding and loading python2.5 dynamically but
I was afraid it would complicate things unnecessarily.