pmxbot IRC Log Viewer

[18:44:51] <wiredfool_> IIRC, my biggest concern with 243 at the time was to not screw up with the licensing, which was addressed.. Secondary was actually making sure that all the ico stuff worked, but it looks good on that front.

[19:07:46] <Arfrever> aclark: One hunk of patch from issue #258, which failed to apply for you, was removing trailing whitespace.

[19:15:59] <aclark> Arfrever: cool thanks

[19:16:02] <aclark> wiredfool_: great

[19:16:12] <aclark> so now we just need cgohlke's thing to work

[19:16:53] <aclark> Bunch of weird stuff with __future__ here: https://travis-ci.org/python-imaging/Pillow/jobs/8599761

[19:21:28] <Arfrever> aclark: You can delete all trailing whitespace.

[19:21:56] <Arfrever> aclark: In top-level directory:

[19:21:56] <Arfrever> find * -type f "-(" -name "*.bdf" -o -name "*.c" -o -name "*.h" -o -name "*.py" -o -name "*.rst" -o -name "*.txt" "-)" -exec sed -e "s/[[:space:]]*$//" -i {} \;

[19:21:56] <Arfrever> sed -e "s/[[:space:]]*$//" -i Sane/CHANGES Sane/README Scripts/README

[19:23:19] <aclark> Arfrever: I'll stick with Line 58 for now, but yeah.

[19:25:53] <Arfrever> aclark: I have reviewed output of `git diff` after these commands and it looks correct.

[19:29:20] <wiredfool_> The namespace pull request has been out there for at least a month and a half, and now we're mucking with it?

[19:29:59] <aclark> wiredfool_: cgholke is apparently

[19:30:08] <wiredfool_> nothing like a deadline

[19:30:41] <aclark> Well, I just poked him to be fair

[19:32:17] <wiredfool_> so, travis has been running on the installed version since the initial travis commit.

[19:32:46] <wiredfool_> I'm not sure why we should be changing it to an inplace test on a release day

[19:32:55] <aclark> Arfrever: find: -(: unknown primary or operator

[19:33:06] <Arfrever> aclark: What operating system?

[19:33:07] <aclark> wiredfool_: because that's how the tests were designed to run apparently

[19:33:11] <aclark> Arfrever: OS X

[19:33:16] <aclark> zsh

[19:33:18] <Arfrever> aclark: It works with GNU find.

[19:34:00] <wiredfool_> aclark: I'd say, if we're going to change the travis setup like that, lets do it early in a release cycle

[19:34:59] <aclark> wiredfool_: why do we care? I need cgohlke to sign off since he's going to build the windows eggs

[19:35:08] <aclark> if there is a clear path to fixing it…

[19:35:23] <Arfrever> aclark: Try:

[19:35:24] <Arfrever> for ext in "*.bdf" "*.c" "*.h" "*.py" "*.rst" "*.txt" ; do find * -type f -name "${ext}" -exec sed -e "s/[[:space:]]*$//" -i {} \; ; done

[19:36:13] <aclark> Arfrever: sed: -i may not be used with stdin

[19:36:18] <aclark> Arfrever: I'll just do it from Linux

[19:36:20] <aclark> give me a min

[19:39:51] <wiredfool_> aclark: whatever. If he can make it work.

[19:41:13] <aclark> well, I need to understand what is going on…

[19:41:35] <aclark> It may be that we wait on his PR, i just don't know yet

[19:41:45] <aclark> But I'd rather get it in and be done and make everyone happy.

[19:42:13] <aclark> wiredfool_: if i can fix the tests you don't have any other issue right?

[19:50:58] <Arfrever> aclark: I noticed that test suite leaves font.png file in top-level directory of working copy.

[19:58:45] <Arfrever> aclark: I filed issue #262.

[20:12:24] <aclark> Arfrever: cool thanks

[23:04:13] <Arfrever> aclark: Re issue #264: Which Python version do you use?

[23:04:33] <aclark> Arfrever: 2.7

[23:05:26] <Arfrever> aclark: Does the problem occur outside of buildout?

[23:06:00] <aclark> Arfrever: not using buildout

[23:07:46] <Arfrever> aclark: Show: bin/python -c 'import sys; print(sys.path)'

[23:09:08] <aclark> Arfrever: https://gist.github.com/aclark4life/5897378

[23:11:40] <Arfrever> aclark: Is __future__.py anywhere?

[23:14:34] <aclark> Arfrever: https://gist.github.com/aclark4life/5897389

[23:15:27] <Arfrever> aclark: Show: bin/python -c 'import __future__; print(__future__)'

[23:15:57] <aclark> Arfrever: <module '__future__' from '/Users/aclark/Developer/buildout.python/parts/opt/lib/python2.7/__future__.pyc'>

[23:17:49] <Arfrever> aclark: Show: bin/python -S -c 'import __future__; print(__future__)'

[23:19:02] <aclark> Arfrever: https://gist.github.com/aclark4life/5897401

[23:19:12] <aclark> Arfrever: Note I can import __future__ just fine normally

[23:19:23] <Arfrever> aclark: -S is appended when --installed is not used.

[23:19:51] <Arfrever> aclark: See lines 22-24 and 50 in Tests/run.py

[23:22:08] <Arfrever> It works for me with -S: $ python2.7 -S -c 'import __future__; print(__future__)'

[23:22:08] <Arfrever> <module '__future__' from '/usr/lib64/python2.7/__future__.pyc'>

[23:23:11] <Arfrever> aclark: Your Python seems to be installed in -S-unfriendly way :) .

[23:23:28] <Arfrever> aclark: Show: bin/python -S -c 'import sys; print(sys.path)'

[23:24:07] <aclark> Arfrever: Ah! Thanks. Same thing on Travis?

[23:24:36] <aclark> Arfrever: https://gist.github.com/aclark4life/5897422

[23:29:27] <Arfrever> aclark: Is this a "framework" installation?

[23:30:13] <aclark> Arfrever: no, buildout.python: https://github.com/collective/buildout.python

[23:32:14] <Arfrever> aclark: Maybe buildout.python places stdlib in a location unknown by core code of Python interpreter. (Python interpreter without -S option imports site.py, which adds some additional directories to sys.path.)

[23:33:41] <aclark> Arfrever: It shouldn't… let me see

[23:34:56] <Arfrever> aclark: In my system installation of Python on GNU/Linux, __future__.py is in location, which is in sys.path even with -S option.

[23:35:15] <aclark> ah

[23:35:39] <Arfrever> aclark: Your __future__.py is /Users/aclark/Developer/buildout.python/parts/opt/lib/python2.7/__future__.py as you previously pasted.

[23:38:03] <aclark> Arfrever: OK here's Ubuntu where I get the same error: https://gist.github.com/aclark4life/5897476

[23:38:20] <Arfrever> Any */buildout.python/* directories are absent in base (i.e. hardcoded in Python interpreter outside of site.py) sys.path in your Python.

[23:39:51] <Arfrever> aclark: Where is __future__.py on Ubuntu? (bin/python -c 'import __future__; print(__future__)')

[23:40:39] <aclark> <module '__future__' from '/usr/lib/python2.7/__future__.pyc'>

[23:42:54] <Arfrever> aclark: Show sys.path with -S and show sys.prefix and sys.exec_prefix.

[23:44:29] <Arfrever> (sys.path on Ubuntu shows mixing of system installation and local installation of Python.)

[23:44:38] <aclark> Ah, ugh

[23:45:12] <Arfrever> It contains some /home/aclark/Developer/Pillow/lib/python2.7/* directories and some /usr/lib/python2.7/* directories.

[23:45:42] <aclark> https://gist.github.com/aclark4life/5897501

[23:48:30] <Arfrever> aclark: Is anything found by?: find /home/aclark/Developer/Pillow/lib/python2.7 -name __future__.py

[23:49:28] <aclark> no

[23:49:32] <Arfrever> aclark: How have you installed local (not system) installation of Python on Ubuntu?

[23:49:54] <aclark> I'm using the system Python on ubuntu

[23:49:59] <aclark> no local python

[23:50:30] <Arfrever> aclark: What is "bin/python" on Ubuntu?

[23:51:08] <Arfrever> aclark: I.e. what is /home/aclark/Developer/Pillow/bin/python

[23:51:50] <aclark> virtualenv

[23:52:56] <Arfrever> aclark: Are any *.py files in /home/aclark/Developer/Pillow/lib/python2.7 ?

[23:54:06] <aclark> https://gist.github.com/aclark4life/5897530

[23:54:38] <Arfrever> aclark: Which version of virtualenv?

[23:55:27] <Arfrever> aclark: I think that virtualenv might be simply incompatible with -S option.

[23:56:53] <Arfrever> aclark: python_options.append("-S") (lin 23) probably could be removed.

[23:57:00] <Arfrever> s/lin/line/

[23:57:21] <aclark> ah

[23:58:42] <aclark> https://gist.github.com/aclark4life/5897552

[23:59:07] <Arfrever> IIRC virtualenv works by creating special site.py (ignored with -S), which hacks many things.

Log file Viewer

Help | Karma | Search:

#pil logs for Sunday the 30th of June, 2013