PMXBOT Log file Viewer

Help | Karma | Search:

#pil logs for Thursday the 31st of July, 2014

(Back to #pil overview) (Back to channel listing) (Animate logs)
[05:59:07] <travis-ci> [travis-ci] wiredfool/Pillow#441 (lcov - 5c2b64f : wiredfool): The build passed.
[05:59:07] <travis-ci> [travis-ci] Change view : https://github.com/wiredfool/Pillow/compare/a4a4d021cc14...5c2b64f3ef53
[05:59:07] <travis-ci> [travis-ci] Build details : http://travis-ci.org/wiredfool/Pillow/builds/31302881
[06:22:28] <travis-ci> [travis-ci] wiredfool/Pillow#442 (lcov - 7e8d583 : wiredfool): The build passed.
[06:22:28] <travis-ci> [travis-ci] Change view : https://github.com/wiredfool/Pillow/compare/5c2b64f3ef53...7e8d583b9621
[06:22:28] <travis-ci> [travis-ci] Build details : http://travis-ci.org/wiredfool/Pillow/builds/31303537
[09:41:15] <qua-non> hey guys
[09:41:37] <qua-non> packaging pillow with pyinstaller gives me the following error
[09:41:49] <qua-non> ImportError: The _imagingft C module is not installed
[09:42:26] <qua-non> I tried with pyinstaller stable and develop
[09:42:42] <qua-non> same issue with both
[09:43:38] <qua-non> log http://paste.opensuse.org/5665274
[16:16:01] <wiredfool> qua-non: I'm guessing that your freetype headers or library weren't found when building.
[16:16:18] <wiredfool> what does python selftest.py say after build
[16:16:28] <qua-non> wiredfool: I figured it out it was a version mismatch
[16:16:41] <wiredfool> ah. figures
[16:16:55] <qua-non> with freetype and imgingft and freetype
[16:17:22] <qua-non> btw pil is pretty fast. the delay I was talking about yesterday was totally due to disk access slowness
[16:17:32] <wiredfool> oh. good to know
[16:17:49] <qua-non> so I can manipulate a big image in under a sec even on android
[16:18:26] <qua-non> besides my app is about batch processing nd right now the bottleneck is diskio and not pil
[16:18:49] <wiredfool> cool
[16:19:10] <qua-non> pillow is very amazing at processing big images. I am very impressed.
[16:20:44] <qua-non> renderscript is very nice, thanks for pointing me to it. However I would rather use OpenGL shaders(if need be) directly as I get cross platform compatibility and speed.
[16:21:31] <qua-non> RenderScript would be a lot more attractive if I could use it one desktops and on ios too.
[16:22:34] <wiredfool> yeah. it is android specific
[16:23:39] <wiredfool> it doesn't feel all that different from other pixel shader type stuff that I've seen around, but I haven't gone and ported anything from it elsewhere
[16:24:05] <qua-non> yes, syntax is pretty similar
[16:24:14] <wiredfool> I was impressed with it when I got an image processing pipeline that I was working on from ~1 sec to 20ms.
[16:25:14] <aclark> !m wiredfool
[16:25:14] <pmxbot> you're doing good work, wiredfool!
[16:25:32] <wiredfool> hey aclark
[16:25:38] <aclark> o.
[16:25:41] <aclark> o/
[16:25:57] <wiredfool> have you seen what's in the pipeline?
[16:26:13] <aclark> wiredfool: no, been tuning out because I had to move recently. What's up?
[16:27:06] <wiredfool> I've figured out how to do test coverage for c code + python code on coveralls
[16:27:13] <wiredfool> https://coveralls.io/builds/1026922
[16:27:29] <aclark> nice!
[16:27:47] <wiredfool> good news, we've now got 14k loc verified covered
[16:27:55] <wiredfool> bad news, we're at 73%
[16:28:03] <aclark> heh
[16:28:07] <wiredfool> this build's better: https://coveralls.io/builds/1026950
[16:28:28] <aclark> you mean when you add the *.c we drop in coverage? If so, yeah better to bite the bullet though.
[16:28:34] <wiredfool> yep
[16:28:49] <aclark> I was looking forward to 80% but I'll get over it :-)
[16:28:55] <wiredfool> though, c coverage is generally going to be lower, since it's hard to hit all the out of memory tests
[16:29:04] <aclark> yeah looks great
[16:29:06] <aclark> right
[16:29:27] <wiredfool> and the average number of times a line is hit is... useless
[16:29:34] <aclark> hah
[16:30:03] <wiredfool> somethings hit 9 million times, some things hit once. average, 9k.
[16:30:29] <aclark> yeah
[16:31:00] <wiredfool> anyway, I'm not sure if we should hit 80% before moving the goalpoasts or not
[16:31:56] <aclark> as long as hugovk knows it's coming I'd say OK to merge. I know he's been trying to get to 80%
[16:37:50] <wiredfool> k
[16:38:10] <wiredfool> I'd rather know what our coverage on C is than be worried about an exact number
[16:38:45] <aclark> right
[16:38:59] <wiredfool> And also, I hate moving. Bought a house so that I didn't have to move anymore.
[16:39:06] <wiredfool> that was 3 moves ago.
[16:40:14] <wiredfool> Working with Eric on the MPO stuff was good. He's a good contributor for pillow
[16:41:29] <aclark> Great! Yeah I forget how I met him but great guy.
[16:41:44] <wiredfool> there's definitely some good coverage info on the c code. We're getting a lot, but edge cases are getting missed.: https://coveralls.io/files/256580642#L1182
[16:41:56] <aclark> And yeah, been renting for 9 years and finally bought a house \o/
[16:42:26] <wiredfool> That's awesome
[16:42:42] <aclark> Hah you bought a house and you *still* had to move
[16:43:00] <wiredfool> yeah. bought before I got married.
[16:43:27] <aclark> Ahhh
[16:43:37] <wiredfool> held on to it for a while, got married, 2 years later, moved to the boonies, in a rental
[16:43:45] <aclark> ooo
[16:43:54] <wiredfool> then later bought. in late 2007
[16:44:31] <aclark> wow that was a sketchy time IIRC… or was that right after the crash.
[16:44:47] <wiredfool> yeah. it was maybe 10% off the peak
[16:44:59] <wiredfool> and a year before the real global financial clusterfuck
[16:45:19] <aclark> right!
[16:45:28] <wiredfool> it was a failure of imagination on my part as to how screwed the country was.
[16:46:28] <aclark> right, "I can't believe so many people can fuck up so much… now I know."
[16:47:07] <wiredfool> yeah.
[16:47:50] <wiredfool> so where is your new house?
[16:48:10] <wiredfool> Mt
[16:50:22] <aclark> wiredfool: I moved from Bethesda, MD to… Bethesda, MD. Just a few miles from where I was renting. I work at NIH and both wife and I wanted to stay here.
[16:50:33] <aclark> $$$ though
[16:50:50] <aclark> Hopefully we'll get it back if/when we sell/move again.
[16:51:10] <aclark> (which I'm not in a hurry to do but you never know)
[16:51:20] <wiredfool> yeah, dc area realestate has been crazy for a long time
[16:51:43] <aclark> True
[16:51:51] <wiredfool> iirc, it and SV were the two fastest recoveries
[16:52:14] <aclark> yep
[16:52:28] <wiredfool> NIH does pubmed, right?
[16:52:31] <aclark> wiredfool: what's in Mt Vernon besides George Washington's house? :-)
[16:52:33] <aclark> Yes
[16:53:02] <aclark> I know a guy that works on it
[16:53:05] <wiredfool> I did some consulting work on that at one point, back in the 2002 era
[16:53:09] <aclark> Oh nice
[16:53:16] <wiredfool> not sure my work went anywhere though
[16:53:30] <aclark> I heard it was nasty, technology wise
[16:54:03] <wiredfool> not much at Mt Vernon, other than the house and suburbs.
[16:54:42] <wiredfool> They had a decent interface for getting things out of the system. Which was kind of rare at the time
[18:06:22] <wiredfool> fg