578584 - Update Python in MozillaBuild to 2.7

Status: RESOLVED FIXED

(Firefox Build System :: MozillaBuild, task)

Description

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

I wonder if it's too late to squeeze in Python 2.7 to MozillaBuild 1.5 RC - it's much much more convenient to have a working os.kill function for Windows when attempting to fuzz.

http://docs.python.org/library/os.html#os.kill

(fwiw, Ted updated to Python 2.6.5 on July 3rd (many thanks for that), and Python 2.7 got released on the same day.)

I don't know of any incompatibilities between the 2.5/2.6.5 and 2.7 series though.

Comment 1

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Attachment: Patch for the other files based on http://hg.mozilla.org/mozilla-build/rev/d4b8b4f97e4e

Binary packages are here:

http://pypi.python.org/packages/2.7/s/setuptools/setuptools-0.6c11-py2.7.egg

http://www.python.org/ftp/python/2.7/python-2.7.msi

Comment 2

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to comment #1)
> Created attachment 457252 [details] [diff] [review]

I have no idea how to test it since I don't have the environment "specific to the paths on the mozillabuild-builder vm" as specified in make-mozillabuild.bat.

Comment 3

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

subprocess.check_output would also be very useful.

http://docs.python.org/library/subprocess.html#subprocess.check_output

Comment 4

[(not currently active) Ted Mielczarek]

I thought Python 2.6 was sufficient for you? I don't have any real opposition here, it just seems like extra churn.

Comment 5

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to comment #4)
> I thought Python 2.6 was sufficient for you? I don't have any real opposition
> here, it just seems like extra churn.

The extra stuff would be regarded as nice-to-have, but it's really ok if it's way too much of a hassle (which I've tried to help as much as possible with the patch)..

Comment 6

[Justin Wood (:Callek)]

imho its probably worth missing this milestone for this milestone of mozbuild, since we already have an rc out; if we need to respin the rc for some reason, then I don't see a reason we can't take this along though.

Comment 7

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to comment #6)
> imho its probably worth missing this milestone for this milestone of mozbuild,
> since we already have an rc out; if we need to respin the rc for some reason,
> then I don't see a reason we can't take this along though.

Armen has another possibly nice-to-have bug to ride-along in bug 579159..

Comment 8

[(not currently active) Ted Mielczarek]

Comment on attachment 457252 [details] [diff] [review]
Patch for the other files based on http://hg.mozilla.org/mozilla-build/rev/d4b8b4f97e4e

Looks fine, but I haven't tested it.

Comment 9

[Mounir Lamouri (:mounir)]

Isn't this patch ready to be pushed?

Comment 10

[(not currently active) Ted Mielczarek]

It needs to be tested, and I never got around to it.

Comment 11

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Attachment: v2 patch for 2.7.1 based on previous patch

untested patch for 2.7.1 based on the previous patch

Comment 12

[Ed Morley [:emorley]]

(In reply to comment #4)
> I thought Python 2.6 was sufficient for you? I don't have any real
> opposition here, it just seems like extra churn.

I tried to get the reviewer chooser prototype working via the Mozilla Build environment, but it requires Python 2.7 + to work, or else I get an |ImportError: cannot import name Counter| error...

http://weblog.latte.ca/blake/employment/mozilla/thunderbird/reviewerChooser

I know I can just install another copy of python elsewhere, but it's just something else to remember to do if I'm using a different machine / then means I've got two versions installed for the sake of it.

Thanks :-)

Comment 15

[Myk Melez [:myk] [@mykmelez]]

Ted: any chance you have time to test Gary's new patch?  Python 2.7 would be handy for an accumulating number of reasons, to judge by the comments here and duplicates filed.

Comment 16

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

Attachment: v3 patch for 2.7.2 based on previous obsoleted 2.7.1 patch (which in turn was based on the original patch)

Here's an untested patch for 2.7.2, edited quickly from the 2.7.1 patch which was edited quickly from the 2.7 one.

Comment 17

[Gary Kwong [:gkw] [:nth10sd] (NOT official MoCo now)]

(In reply to Gary Kwong [:nth10sd] from comment #16)
> Created attachment 551651 [details] [diff] [review] [diff] [details] [review]
> v3 patch for 2.7.2 based on previous obsoleted 2.7.1 patch (which in turn
> was based on the original patch)
> 
> Here's an untested patch for 2.7.2, edited quickly from the 2.7.1 patch
> which was edited quickly from the 2.7 one.

Binary packages are here:

http://pypi.python.org/packages/2.7/s/setuptools/setuptools-0.6c11-py2.7.egg

http://www.python.org/ftp/python/2.7.2/python-2.7.2.msi

Comment 18

[(not currently active) Ted Mielczarek]

Testing the patch is just as much of a pain as building a new version of MozillaBuild. I'd like to do it, but it's very time-consuming. khuey and I have been thinking about doing a new MozillaBuild release for a while, we just haven't found time.

Comment 19

[Jesse Ruderman]

Note to self: don't panic, bug 659881 is mac-specific and we have a workaround.

Comment 20

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

I built mozilla-build with this and didn't see anything obviously wrong.

http://hg.mozilla.org/mozilla-build/rev/240a0e9e6e7e