#pypy on 2019-02-23 — irc logs at freenode.irclog.whitequark.org

2018-11-25 07:18 arigato changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://quodlibet.duckdns.org/irc/pypy/latest.log.html#irc-end ) | use cffi for calling C | mac OS and Fedora are not Windows

00:42 kipras has quit [Ping timeout: 246 seconds]

00:56 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6304 [default]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/1009 [default]

01:49 <kenaan> rlamy py3.6 ed559ffaf9d1 /pypy/tool/pytest/apptest.py: Ignore _cffi_backend when checking for usemodules in -A tests

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4441 [py3.6]

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6305 [py3.6]

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/5077 [py3.6]

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4249 [py3.6]

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6045 [py3.6]

02:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7316 [py3.6]

02:08 <bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1982 [default]

02:14 speeder39_ has quit [Quit: Connection closed for inactivity]

02:32 moei has quit [Ping timeout: 258 seconds]

02:33 moei has joined #pypy

02:47 _whitelogger has joined #pypy

02:47 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/5077 [py3.6]

02:52 jcea has quit [Ping timeout: 264 seconds]

03:28 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6045 [py3.6]

03:35 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/6305 [py3.6]

03:45 marky1991 has joined #pypy

03:49 marky1991 has quit [Ping timeout: 246 seconds]

04:22 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/1009 [default]

04:47 egregius313 has joined #pypy

04:51 egregius313 has quit [Client Quit]

05:00 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4441 [py3.6]

05:01 egregius313 has joined #pypy

05:03 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7316 [py3.6]

05:04 egregius313 has quit [Client Quit]

05:23 forgottenone has joined #pypy

05:30 xcm has quit [Remote host closed the connection]

05:32 xcm has joined #pypy

05:43 speeder39_ has joined #pypy

05:57 jacob22 has quit [Read error: Connection reset by peer]

05:57 bivab has joined #pypy

06:00 jacob22 has joined #pypy

07:25 <fijal> EWDurbin: yeah I manage pypy DNS

07:25 <fijal> I think

07:30 tazle_ has joined #pypy

07:30 fryguybo1 has joined #pypy

07:31 bogner_ has joined #pypy

07:40 __peke__ has quit [Excess Flood]

07:40 dmalcolm has quit [Excess Flood]

07:40 kanaka has quit [Excess Flood]

07:40 tazle has quit [Write error: Broken pipe]

07:40 fryguybob has quit [Remote host closed the connection]

07:40 bogner has quit [Remote host closed the connection]

07:40 __peke__ has joined #pypy

07:41 dmalcolm has joined #pypy

07:41 kanaka has joined #pypy

07:42 Ai9zO5AP has quit [Ping timeout: 249 seconds]

07:42 pepesza has quit [Ping timeout: 249 seconds]

07:45 pepesza has joined #pypy

07:47 dddddd has joined #pypy

07:56 Ai9zO5AP has joined #pypy

08:22 speeder39_ has quit [Quit: Connection closed for inactivity]

08:35 <mattip> do we have a better xml story than “pip install lxml” as suggested on the wiki?

08:35 <mattip> https://bitbucket.org/pypy/compatibility/wiki/lxml

08:36 <mattip> the wiki entry mentions lxml-cffi but then discourages trying it

08:54 forgottenone has quit [Ping timeout: 245 seconds]

09:52 dddddd has quit [Remote host closed the connection]

10:08 jacob22 has quit [Read error: Connection reset by peer]

10:12 jacob22 has joined #pypy

10:27 k1nd0f has joined #pypy

10:31 Zaab1t has joined #pypy

10:31 forgottenone has joined #pypy

10:33 xcm has quit [Remote host closed the connection]

10:35 xcm has joined #pypy

10:52 tazle_ has quit [Quit: leaving]

10:52 tazle has joined #pypy

11:30 k1nd0f has quit [Ping timeout: 255 seconds]

12:04 Arfrever has quit [Ping timeout: 245 seconds]

12:39 lritter has joined #pypy

12:43 <nanonyme> Yeah, I don't think that's actively maintained

12:43 <nanonyme> Or?

12:48 logan` has joined #pypy

12:49 kipras has joined #pypy

12:53 <nanonyme> mattip, on a I think theoretical level there was talk of CFFI backend for Cython which might possibly make the lxml story less painful

12:53 logan` has quit [Client Quit]

12:53 logan` has joined #pypy

13:07 kipras has quit [Read error: Connection reset by peer]

13:30 dddddd has joined #pypy

14:11 k1nd0f has joined #pypy

14:13 antocuni has joined #pypy

14:19 jacob22 has quit [Read error: Connection reset by peer]

14:23 jacob22 has joined #pypy

14:33 danieljabailey has joined #pypy

14:36 i9zO5AP has joined #pypy

14:38 Ai9zO5AP has quit [Ping timeout: 255 seconds]

14:50 jcea has joined #pypy

15:23 egregius313 has joined #pypy

15:32 egregius313 has quit [Ping timeout: 268 seconds]

15:34 logan` has quit []

16:04 k1nd0f has quit [Ping timeout: 268 seconds]

16:16 <mattip> nanonyme: it seems then that lxml might be a good test case for that path, it is widely used and no good alternative exists as far as I can tell

16:27 i9zO5AP has quit [Quit: WeeChat 2.3]

16:28 Ai9zO5AP has joined #pypy

16:30 forgottenone has quit [Ping timeout: 255 seconds]

16:47 <nanonyme> mattip, the main issue is apparently that lxml author is also the Cython author so lxml will likely stay as it is

16:51 <nanonyme> But it might be possible to modify Cython to have alternative generation so instead of generating cpyext code it would generate something usable with CFFI

16:59 <mattip> nanonyme: I think that actually makes the case more compelling, assuming the author is willing to cooperate

17:02 <nanonyme> mattip, I don't know enough to help personally, just echoing the conversation

17:22 Arfrever has joined #pypy

17:33 danieljabailey has quit [Ping timeout: 245 seconds]

17:35 forgottenone has joined #pypy

17:38 <mattip> I remounted the /tmp directory for the 32bit chroot on bencher4 so it does not share with the 64bit chroot,

17:39 <mattip> maybe that will solve some of the test crashes

17:39 <mattip> dunno if there are other shared resource that might be causing conflicts: socket ports?

18:16 inhahe_ has quit [Read error: Connection reset by peer]

18:16 inhahe has joined #pypy

18:31 antocuni has quit [Ping timeout: 268 seconds]

19:06 speeder39_ has joined #pypy

20:04 Garen has quit [Read error: Connection reset by peer]

20:04 Garen has joined #pypy

20:06 lritter has quit [Quit: Leaving]

20:09 forgottenone has quit [Quit: Konversation terminated!]

20:20 antocuni has joined #pypy

20:34 nunatak has joined #pypy

20:35 nunatak has quit [Client Quit]

21:26 Ai9zO5AP has quit [Ping timeout: 245 seconds]

21:35 danieljabailey has joined #pypy

21:36 speeder39_ has quit [Quit: Connection closed for inactivity]

21:49 Zaab1t has quit [Quit: bye bye friends]

21:54 k1nd0f has joined #pypy

21:57 <mattip> I tried to encourage the "Talk Python To Me" host to push PyPy, he replied that he tried it for his website and it was really slow

21:57 <mattip> http://disq.us/p/1zylgph

21:58 <mattip> using pyramid and MongoEngine via uWSGI

22:03 antocuni has quit [Ping timeout: 255 seconds]

22:09 <mattip> which is kind of greek to me but maybe someone could give me a pointer to help try again with a more successful stack

22:13 <fijal> mattip: my immediate response would be "how many requests does he get"?

22:16 rubdos has quit [Ping timeout: 244 seconds]

22:23 danieljabailey has quit [Ping timeout: 245 seconds]

22:26 <mattip> fijal: probably not many, but he benchmarked pypy and it came up much slower than cpython, so he now doesn't recommend pypy

22:27 <fijal> well, if he ran it on 10 requests (or even 1000) it's hardly a surprise

22:28 <mattip> 5 times slower, with more memory use

22:28 <EWDurbin> howdy fijal and mattip! y'all should be able to log into the new speed.pypy.org server and look around, i moved the keys you were using for the previous instance over: https://github.com/python/psf-salt/commit/b7cb6d79424d8d8e38513f08d532c82316f4a20b#diff-267d2cc0b5c1a03d40536b1e739b4ee7R327

22:29 <EWDurbin> codespeed.nyc1.psf.io

22:29 <fijal> EWDurbin: thanks!

22:29 <fijal> mattip: as I said, it's really impossible to comment without any more data

22:29 <EWDurbin> it's managed by salt: https://github.com/python/psf-salt/tree/18.04/salt/codespeed

22:29 <fijal> it can *very easily* come 5x slower with more memory usage, sure

22:29 <EWDurbin> and the source is now based on https://github.com/python/codespeed/tree/speed.pypy.org

22:30 <EWDurbin> would you like to be added as collaborators on that repo?

22:30 <EWDurbin> also, whenever you get a chance to update DNS, speed.pypy.org CNAME codespeed-pypy.map.psf.io will do the trick

22:31 <mattip> EWDurbin: thanks. Trying https://codespeed.nyc1.psf.io, it isn't responding right now

22:32 <EWDurbin> mattip: ssh access only. you can preview by fiddling your etc hosts

22:32 <EWDurbin> to point at....

22:32 <mattip> ok

22:32 <EWDurbin> 45.55.99.72

22:35 <mattip> cool, it works :)

22:36 <mattip> it's not your fault, this happened on the old site too, but if you open a dev console and look at the network calls, loading

22:36 <mattip> https://speed.pypy.org/comparison/

22:37 k1nd0f has quit [Ping timeout: 250 seconds]

22:37 <mattip> gives a 502 on the get for https://speed.pypy.org/comparison/json/?_=1550961295130

22:38 <EWDurbin> it's as close as was possible to how the old site was deployed, so i'm not surprised it didn't magically fix anything :)

22:38 <mattip> ok, thanks

22:38 <EWDurbin> slow routes are slow

22:39 <EWDurbin> FWIW it desperately needs some love... it's not compatiable with Django > 1.5 (i tried)

22:39 <mattip> :(

22:39 <kenaan> rlamy default 5f79cbf51fda /pypy/module/cpyext/: Correctly initialize the d_type and d_name members of builtin descriptors. This fixes a segfault related to classm...

22:39 <mattip> so yes, please give us commit rights to the repo

22:40 <kenaan> rlamy py3.6 26a7e7f30d93 /pypy/module/cpyext/: hg merge default

22:40 <EWDurbin> mattip: GH user?

22:40 <EWDurbin> and fijal?

22:40 <mattip> mattip and fijal

22:40 rubdos has joined #pypy

22:40 <mattip> easy

22:40 <EWDurbin> :-D

22:41 rubdos has quit [Read error: Connection reset by peer]

22:42 <mattip> nice that we have https there now

22:42 <EWDurbin> i ain't no goof, TLS all the things

22:42 <EWDurbin> :)