cfbolz 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 | if a pep adds a mere 25-30 [C-API] functions or so, it's a drop in the ocean (cough) - Armin
BPL has quit [Quit: Leaving]
tsaka__ has quit [Ping timeout: 276 seconds]
dddddd has quit [Remote host closed the connection]
jcea has quit [Quit: jcea]
speeder39_ has joined #pypy
xcm has quit [Read error: Connection reset by peer]
xcm has joined #pypy
_whitelogger has joined #pypy
oberstet has joined #pypy
<mattip> just found darling (wine-like api layer to run darwin code on linux), I wonder if it is good enough to test the macos _ssl code
speeder39_ has quit [Quit: Connection closed for inactivity]
tsaka__ has joined #pypy
<marmoute> fun
<mattip> it uses a kernel module which has this warning "Be prepared for kernel hangups and crashes, and run Darling on a virtual machine if possible"
<mattip> yay
inhahe has quit [Ping timeout: 245 seconds]
inhahe has joined #pypy
<cfbolz> mattip: hey!
<cfbolz> Did you get home OK?
<mattip> yup
<mattip> sprint was productive: I added two lines to hg-git (and a test)
<mattip> that's about 1 line for every 16 hours or so?
<cfbolz> :-)
inhahe has quit [Ping timeout: 276 seconds]
inhahe has joined #pypy
tsaka__ has quit [Ping timeout: 276 seconds]
<marmoute> mattip: well, you also added a test. So that is much more than 2 lines
<mattip> marmoute: not complaining (much), it is just the way things work
<marmoute> mattip: I know, you know, we know
_whitelogger has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
lesshaste has quit [Ping timeout: 245 seconds]
ronan__ has joined #pypy
ronan__ has quit [Client Quit]
ronan__ has joined #pypy
ronan__ has quit [Client Quit]
ronan__ has joined #pypy
ronan__ has quit [Client Quit]
antocuni has joined #pypy
<mattip> tos9: can you use any of the downloads on buildbot.pypy.org/nightly/trunk or buildbot.pypy.org/nightly/py3.6?
<mattip> maybe they require some brew install to run?
<cfbolz> arigato: ping? for some random reason I am again looking at rutf8.py, I am wondering, isn't each element of UTF8_INDEX_STORAGE going to use up 128 bytes of memory due to alignment, even though it only stores 80 bytes of content?
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
tsaka__ has joined #pypy
antocuni has quit [Ping timeout: 246 seconds]
dddddd has joined #pypy
<mattip> tos9: I think we always want to rebuild when packaging. This might have been the source of the linking problem above
<mattip> during translation we build without `embed-dependencies`, then package that _ssl module :(
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4579 [mattip: force build, openssl-for-macos]
<ronan> mattip: I updated https://gist.github.com/rlamy/53712623ea16c9c920c574dd2266dbef so that it reproduces the deadlock every time
<ronan> the issue also affects all current CPython versions, so I'll report upstream
<mattip> cool. Do we have a mitigation (other than skipping the test :) )
<mattip> nice detective work
<ronan> no, I don't know how to fix or mitigate the issue
<ronan> I don't know why it's only visible on pypy either
<ronan> but maybe it's something about the thread switching interval?
<mattip> ronan: maybe I am doing something wrong. I run the script with cpython and it prints an exception, does not hang
<ronan> mattip: what exception?
<mattip> it only prints the "start thread", "end thread" twice
<mattip> ModuleNotFoundError: No module named 'package'
<ronan> huh, weird
<ronan> I guess it means that has_deadlock() works correctly for you
<ronan> which version are you running it on?
<mattip> cpython 3.6.8, pypy3 d21e9a6b4037 from Aug 12 2019
<mattip> on ubuntu 18.04
<ronan> I'm on Ubuntu 16.04
<marmoute> mattip: looks like gitlab already has some bitbucket import capability
<marmoute> so I assume importing the issue tracker in heptapod will be "simple"
<mattip> marmoute: maybe. What would the architecture look like? Can we use the public gitlab.com instance?
<mattip> (it seems the pypy group is already taken on gitlab.com :(
<ronan> mattip: I updated the gist with a lower-level test_deadlock.py, does it hang for you?
<mattip> ronan: pypy3 deadlocks after printing T1: 1 (and my cpu fan starts churning)
<mattip> as does cpython3
<ronan> OK, it's the same for me
<marmoute> mattip: if you are using heptapod. that won't be on the main gitlab.com instance.
<mattip> just trying to understand. So where would it be hosted?
<marmoute> a dedicated instance.
antocuni has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4579 [mattip: force build, openssl-for-macos]
marky1991 has joined #pypy
Dejan has joined #pypy
<Dejan> Hi guys
<Dejan> I was just looking at http://buildbot.pypy.org/nightly/py3.6/ - what does that first number in "own tests" mean?
<Dejan> it jumped from 7852 to 15698 for linux64 builds
<ronan> Dejan: it's how many tests ran
<ronan> sometimes it aggregates several runs
<Dejan> aham, that explains it
<Dejan> thanks ronan
<Dejan> 5-6 failures seems like a good progress...
<Dejan> (own tests)
Taggnostr has quit [Remote host closed the connection]
Taggnostr has joined #pypy
Rhy0lite has joined #pypy
jcea has joined #pypy
<mattip> there is a survey about the release cycle for cpython (PEP 602) https://twitter.com/llanga/status/1171426661893435397
<mattip> here is my response
<Dejan> bugfix releases released every month for 12 months (12 bugfix releases) seems overly optimistic to me
<Dejan> does PSF have resources for that?
<mattip> so much noise for so little gain
<Dejan> IMHO bugfix releases could be released every 3-4 months
<Dejan> but yea, if they have resources for every month... then why not :)
jcea has quit [Ping timeout: 264 seconds]
<mattip> because they are at the head of a whip, when they "crack" it the entire ecosystem has to respond
<mattip> by the time it gets to users, they have to update all the time, or start ignoring the bugfixes
<simpson> Yeah, I wonder how they see this interacting with people who want to code to feature subsets.
<mattip> "they" have varied opinions, probably few are aware of that use case
<ronan> hmm, test_threaded_import appears to fail for reasons similar to test_import
<ronan> I'm starting to feel that _ModuleLock is just broken
<ronan> mattip: fixing test_import might not be so urgent, considering every existing Python3 interpreter has the same issue
marky1991 has quit [Read error: Connection reset by peer]
marky1991 has joined #pypy
lesshaste has joined #pypy
lesshaste is now known as Guest84347
<tos9> mattip: I can try those -- is there a reason they should work?
<tos9> mattip: The branch we pushed didn't get merged to trunk / py3.6, did it?
* tos9 downloads nightly
<mattip> no, I tried pushing a fix to activate your rebuild option but something is broken
<tos9> Ah, hm. OK.
<tos9> (Sorry, waiting for my mercurial to update... :/)
<mattip> I think I see the problem, in 89c76ee0eafd you removed ['--without-openssldir'], but should have left an empty list?
Guest84347 has quit [Remote host closed the connection]
<mattip> tos9: it turns out I can use "darling" to run macos binaries on my linux machine. I can't compile, but I can run pypy
* mattip pushed a fix
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4580 [mattip: force build, openssl-for-macos]
<tos9> Aha cool thanks for the catch
<tos9> And I have no idea what darling is but that sounds nice
<tos9> (Finishing up a meeting but will be back again in a bit...)
jcea has joined #pypy
<mattip> ronan: agreed it is not that important. Can we find a way to decrease the timeout killer for this test so failure does not delay the build?
oberstet has quit [Remote host closed the connection]
<ronan> mattip: I think the test is supposed to shut itself down after 15 minutes, but it hangs at shutdown
waldhar has quit [Quit: ZNC 1.7.4 - https://znc.in]
antocuni has quit [Ping timeout: 240 seconds]
jcea has quit [Ping timeout: 276 seconds]
antocuni has joined #pypy
<ronan> pfff... trying to figure out why test_threaded_import consistently fails on pypy but not on CPython
<mattip> tos9: got a little further, but still no luck
<mattip> let's try once more
xcm has quit [Remote host closed the connection]
<bbot2> Exception: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4580 [mattip: force build, openssl-for-macos]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4581 [mattip: force build, openssl-for-macos]
xcm has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
antocuni has quit [Ping timeout: 276 seconds]
waldhar has joined #pypy
<ronan> Yay! CPython's IMPORT_NAME has an "optimisation" that behaves differently from __import__()....
<mattip> ronan: so they _would_ get the deadlock, but avoid it by having a specific bytecode?
<ronan> yes, I just checked that they fail like pypy if you replace 'import A' with 'importlib.import_module('A')'
<tos9> mattip: cool, that error looks more promising
<tos9> at least we're failing now presumably in the difference between how you compile libressl and openssl
<mattip> tos9: this thing is a bit unweildy, could you run it locally?
<mattip> it seems to be pretty straightforward to reproduce the steps in _build_dependency, and make sure they still work for building gdbm as well
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4581 [mattip: force build, openssl-for-macos]
ssbr` has joined #pypy
Rhy0lite has quit [Quit: Leaving]
<mattip> tos9: the _ssl build error reproduces on linux. good.
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4582 [mattip: force build, openssl-for-macos]
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
CrazyPython has joined #pypy
tsaka__ has quit [Ping timeout: 258 seconds]
CrazyPython has quit [Read error: Connection reset by peer]
kipras has joined #pypy
xcm has quit [Remote host closed the connection]
kipras has quit [Read error: Connection reset by peer]
xcm has joined #pypy
speeder39_ has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/4582 [mattip: force build, openssl-for-macos]
<tos9> mattip: semi-back... is there something for me to run still or did you reproduce
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
jacob22 has quit [Quit: Konversation terminated!]
jacob22 has joined #pypy
jacob22 has quit [Client Quit]
jacob22 has joined #pypy
tsaka__ has joined #pypy
jacob22 has quit [Quit: Konversation terminated!]
jacob22 has joined #pypy
jacob22 has quit [Quit: Konversation terminated!]
jacob22 has joined #pypy
jcea has joined #pypy
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
marky1991 has quit [Ping timeout: 240 seconds]