01:19
xcm has quit [Read error: Connection reset by peer]
01:20
xcm has joined #pypy
02:23
lritter has quit [Ping timeout: 268 seconds]
03:34
jcea has quit [Ping timeout: 265 seconds]
07:05
jvesely has quit [Ping timeout: 268 seconds]
08:26
_whitelogger has joined #pypy
09:14
dddddd has quit [Ping timeout: 268 seconds]
10:32
glyph has quit [Quit: End of line.]
10:32
glyph has joined #pypy
11:25
<
cfbolz >
mattip: yes, 2.7 does not do the putenv check
12:30
<
mattip >
ok. I guess we should test and fix that too, since cpython2.7 rejects '=' in the key
12:32
<
cfbolz >
mattip: weird:
12:32
<
mattip >
i found who is doing "import os". open goes to _io.open which goes through interp_textio._determine_encoding
12:32
<
mattip >
which not only import app-level os but imports app-level _bootlocale as well
12:32
<
mattip >
isn't that slowing down our use of open quite a bit?
12:34
<
mattip >
cfbolz: in what sense is a '=' in an environment variable "illegal"?
12:35
<
cfbolz >
mattip: in the sense of "python3 rejects it"
12:45
xcm has quit [Remote host closed the connection]
12:47
xcm has joined #pypy
13:08
dddddd has joined #pypy
14:07
jcea has joined #pypy
14:12
jcea has quit [Ping timeout: 248 seconds]
14:23
jcea has joined #pypy
14:44
<
mattip >
made a bunch of changes to test_posix2
14:44
lritter has joined #pypy
14:44
xcm has quit [Remote host closed the connection]
14:45
kenaan has joined #pypy
14:45
<
kenaan >
mattip default 72aff1d1e6fa /pypy/module/posix/test/test_posix2.py: backport some of the changes from py3.6
14:45
<
kenaan >
mattip default 03da8b4056eb /pypy/module/posix/test/test_posix2.py: do not import stdlib os
14:45
<
kenaan >
mattip default 2821b0633b53 /pypy/: simplify the call to wrap_oserror2
14:45
<
kenaan >
mattip default 30cf519fd69a /pypy/: pep-8 cleanup
14:45
<
kenaan >
mattip default c14832883def /pypy/module/posix/test/test_posix2.py: fix failing win32 test
14:45
xcm has joined #pypy
14:49
kcr has joined #pypy
14:50
Alex_Gaynor has quit []
14:50
Alex_Gaynor has joined #pypy
14:55
<
mattip >
uhhh, something went wrong there
15:04
<
mattip >
how do I back out all what I merged to default after c14832883def?
15:04
<
mattip >
somehow that all went wrong
15:21
xcm has quit [Ping timeout: 260 seconds]
15:25
xcm has joined #pypy
15:51
<
mattip >
I have a revert commit but that means all the hg history is going to be messed up
15:53
<
mattip >
I guess the other option is to force-push
15:54
xcm has quit [Read error: Connection reset by peer]
15:56
xcm has joined #pypy
15:57
<
cfbolz >
mattip: please don't force push
15:57
<
cfbolz >
having slightly messy history is ok
15:57
<
mattip >
ok. I mistakenly somehow pushed a merge of py3.6 into default instead merge of default into py3.6
15:58
<
mattip >
so it means hg blame will have a glitch
16:00
<
mattip >
cfbolz: other than a big revert commit, any ideas how to get back to commit c14832883def?
16:02
<
ronan >
mattip: how about just dropping everything after c14832883def?
16:02
<
mattip >
that's fine, but it is a force push, no?
16:03
<
ronan >
well, it's possibly a strip
16:04
<
mattip >
cfbolz: thoughts?
16:05
<
cfbolz >
Hmmm, blame being wrong is indeed annoying :-(
16:06
<
cfbolz >
I don't know
16:06
<
mattip >
it's from my last push, maybe only very few people have pulled?
16:08
rguillebert has quit []
16:08
rguillebert has joined #pypy
16:09
<
cfbolz >
mattip: likely even zero, today
16:11
<
mattip >
ronan: how would that even work with bitbucket? How can I tell it "forget these commits" ?
16:11
<
ronan >
there's a tab in the admin settings
16:13
<
mattip >
ok, trying to "unpublish"
16:14
<
ronan >
mattip: unpublishing isn't really useful, I think
16:14
<
ronan >
it just changes the phase to draft
16:15
fijal has joined #pypy
16:16
<
mattip >
yes, now trying a "strip"
16:17
<
mattip >
hopefully no-one pulled those changes
16:48
jvesely has joined #pypy
17:15
<
cfbolz >
mattip: did it work?
17:33
i9zO5AP has joined #pypy
17:34
Ai9zO5AP_ has joined #pypy
17:35
Ai9zO5AP has quit [Ping timeout: 260 seconds]
17:38
i9zO5AP has quit [Ping timeout: 240 seconds]
17:49
<
kenaan >
mattip py3.6 eee8317e40c5 /: move test to extra_tests, speed up unstranslated tests by 50%
17:49
<
kenaan >
mattip py3.6 3b12b05e7a25 /: redo tests to not import app-level lib-python/3/os.py
17:49
<
kenaan >
mattip py3.6 8fb4dc6f07e6 /pypy/module/: fix for path-as-memoryview on win32, avoid app-level code in _io.open
17:49
<
kenaan >
mattip py3.6 c7efe1e62575 /pypy/interpreter/error.py: emit correct exception class
17:49
<
kenaan >
mattip py3.6 4611cc30e74d /pypy/module/posix/test/test_posix2.py: fix typos, also posix.environ returns bytes
17:49
<
kenaan >
cfbolz py3.6 23ce279ad084 /pypy/module/posix/interp_posix.py: make putenv complain about bad names (test already existed) (grafted from 89255ed57e9fa884c8bbeb50f19f2194b4fe698f)
17:55
<
mattip >
it worked, the question is if anyone pulled from the bitbucket repo and got those changes
17:56
<
mattip >
I should have done this work on a branch, but it bounces between default and py3.6 so I would have needed two branches
17:56
<
mattip >
next time ...
17:57
xcm has quit [Read error: Connection reset by peer]
17:58
ammar2 has quit [Ping timeout: 245 seconds]
17:58
xcm has joined #pypy
18:05
<
kenaan >
mattip default 0395c16d6bbe /pypy/module/_io/interp_textio.py: backport part of 8fb4dc6f07e6
18:07
<
mattip >
I wonder if 0395c16d6bbe has performance impact
18:07
<
mattip >
it should, since it avoids app-level code
18:08
<
mattip >
but maybe it is negligible
18:21
ammar2 has joined #pypy
18:25
lauren has joined #pypy
18:30
xcm has quit [Remote host closed the connection]
18:31
xcm has joined #pypy
20:17
ctismer has quit []
20:17
ctismer has joined #pypy
20:41
<
kenaan >
mattip py3.6 0ce9e1439a97 /pypy/module/posix/interp_posix.py: revert bogus code
21:03
<
arigato >
mattip: just in case, you can
21:04
<
arigato >
also fix very bogus checkins like this:
21:05
<
arigato >
if there is a head of "default" to forget, you close the default branch "hg ci --close-branch", and then you go back to the earlier commit on default, and you make a dummy commit from there
21:05
<
arigato >
there are two heads of "default" but one of them is closed and the other is newer, so everything is fine again
22:13
CrazyPython has joined #pypy
22:59
mwhudson has joined #pypy
22:59
mwhudson has joined #pypy
22:59
mwhudson has quit [Changing host]
23:12
xcm has quit [Remote host closed the connection]
23:14
xcm has joined #pypy
23:47
kbtr has joined #pypy
23:49
wleslie has quit [Quit: ~~~ Crash in JIT!]
23:55
CrazyPython has quit [Read error: Connection reset by peer]