cfbolz changed the topic of #pypy to: #pypy PyPy, the flexible snake https://pypy.org | IRC logs: https://quodlibet.duckdns.org/irc/pypy/latest.log.html#irc-end and https://libera.irclog.whitequark.org/pypy | insert pithy quote here
gef has quit [Quit: No Ping reply in 180 seconds.]
gef has joined #pypy
gef has quit [Ping timeout: 250 seconds]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-aarch64/builds/1026 [mattip: force build, sysconfigdata]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/5604 [mattip: force build, sysconfigdata]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7713 [mattip: force build, sysconfigdata]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-64/builds/275 [mattip: force build, sysconfigdata]
gef has joined #pypy
<cfbolz> Morning
<mattip> hi
<cfbolz> something is still wrong with testcapimodule
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7713 [mattip: force build, sysconfigdata]
<mattip> when importing it I get "undefined symbol: Py_BuildValue"
<mattip> which should be PyPy_BuildValue
<mattip> my bad, the c-extension module is not built?
<mattip> right, confusion, it is built in a temp directory, then imported, then copied to the final location
<mattip> my first comment was correct: it is not getting the proper include paths or something
<mattip> on line 6532 _testcapimodule.c undefines Py_BuildValue which breaks on PyPy
<mattip> I found this by copying the compile command for _testcapimodule.c which is printed during the build stage of importing _testcapi
<mattip> and adding a "-E" so it only precompiles the file, then looking for "\<Py_BuildValues\>"
<mattip> usually, Py_BuildValue is defined as _Py_BuildValue_SizeT, this test wants to actually use Py_BuildValue (without PY_SSIZE_T_CLEAN)
<mattip> which is defined in Python/modsupport.c as calling va_build_value(format, va, 0) instead of va_build_value(format, va, FLAG_SIZE_T)
<mattip> which we copy verbatum into pypy
gef has quit [Ping timeout: 252 seconds]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/5604 [mattip: force build, sysconfigdata]
gef has joined #pypy
<mattip> pushed a fix, you can apply locally to the c file in 3c6353adb043
<cfbolz> mattip: thanks! that would have taken me ages :-(
<cfbolz> cool, now I can run test_call
<cfbolz> of course stuff fails 😅
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-64/builds/275 [mattip: force build, sysconfigdata]
Julian has joined #pypy
Julian has quit [Ping timeout: 264 seconds]
Eclipse2212 has joined #pypy
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7714 [Carl Friedrich Bolz-Tereick: force build, py3.8-vectorcall]
Julian has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-aarch64/builds/1026 [mattip: force build, sysconfigdata]
gef_ has joined #pypy
gef has quit [Ping timeout: 264 seconds]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7714 [Carl Friedrich Bolz-Tereick: force build, py3.8-vectorcall]
Dejan has quit [Quit: Leaving]
Dejan has joined #pypy
gef_ has quit [Ping timeout: 258 seconds]
atomizer has joined #pypy
Eclipse2212 has quit [Ping timeout: 264 seconds]
Julian has quit [Ping timeout: 252 seconds]
atomizer has quit [Ping timeout: 258 seconds]
* mattip back
<mattip> cfbolz: feel free to ping/share tracebacks if you are stuck
<Dejan> https://foss.heptapod.net/pypy/pypy/-/issues/3386 this solution works... if that is the right thing to do, could someone put it there so next 3.8 build has that fix
Eclipse2212 has joined #pypy
<cfbolz> mattip: I was wondering why the lib-python tests don't run, but it's for a stupid reason
<mattip> right, there is a missing Repr
<cfbolz> mattip: is that from the stdlib merge maybe? the lib-python tests used to run
* cfbolz just adds stuff
<mattip> I think there is a new file
<mattip> Dejan: IMO that is the wrong hack: it would be better to add sys._base_executable
<cfbolz> mattip: no, the file is removed on 3.8
<cfbolz> but I brought it back accidentally via a 3.7 merge at some point
stkrdknmibalz has quit [Quit: WeeChat 3.0.1]
<mattip> ahh
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7715 [Carl Friedrich Bolz-Tereick: force build, py3.8-vectorcall]
Eclipse2212 has quit [Ping timeout: 252 seconds]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/7715 [Carl Friedrich Bolz-Tereick: force build, py3.8-vectorcall]
Eclipse2212 has joined #pypy
<mattip> cfbolz: it seems lib_pypy/_testmultiphase.c does not compile
<cfbolz> Ah
<mattip> it would be nice to solve that separately from vectorcall somehow
<mattip> one strategy would be to merge vectorcall as-is, with the rest of it missing?
<cfbolz> mattip: there's not so much missing
<mattip> just that offset flag thing
<cfbolz> mattip: no, I copy pasted those. But seems some details are still wrong
<mattip> ahh
<mattip> maybe leave it for a follow-up PR?
<cfbolz> mattip: works for me
<mattip> ok, most of the _testmultiphase.c failure was a bad merge
<mattip> but we seem to need PyState_{Find,Add,Remove}Module
<mattip> which, strangely, are interfaces for non multi-phase initialization
<bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/8725 [mattip: force build, py3.8]
<cfbolz> mattip: yes, I removed such a call in testcapi too
Julian has joined #pypy
Julian has quit [Ping timeout: 258 seconds]
Julian has joined #pypy
jryans has quit [Write error: Connection reset by peer]
daubers has quit [Remote host closed the connection]
daubers has joined #pypy
jryans has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/8725 [mattip: force build, py3.8]
Eclipse2212 has quit [Quit: WeeChat 3.1]
Julian has quit [Ping timeout: 264 seconds]
Julian has joined #pypy
Dejan has quit [Quit: Leaving]
habnabit_ has quit [*.net *.split]
idnar has quit [*.net *.split]
idnar_ has joined #pypy
habnabit_ has joined #pypy
habnabit_ has joined #pypy
habnabit_ has quit [Changing host]
Julian has quit [Quit: leaving]
<mattip> cool, the lib-python test reporting on py3.8 now works.
<mattip> Parsing the FAILED lines in the stdout https://paste.ubuntu.com/p/WZVfNQNpzW/
<mattip> yeilds {'failures': 484, 'errors': 770, 'skipped': 1491}
<cfbolz> mattip: and how many passed?
<mattip> hmm
<mattip> total was 11013 tests, so ~10000
<cfbolz> cool
<mattip> also, this one liner sorts the error messages by exception + text
<mattip> grep '^[A-Za-z]*Error' /tmp/stdout.txt |sort |uniq -c |sort -n
<mattip> and shows
<mattip> 280 AttributeError: module 'signal' has no attribute 'valid_signals'
<mattip> 197 RuntimeError: no running event loop
<mattip> 98 ModuleNotFoundError: No module named '_posixshmem'
<cfbolz> mattip: right, the second is top level await, the third multiprocessing stuff
<mattip> that grep catches 1386 lines, with a long tail of unique errors
<cfbolz> right
<mattip> cool. So maybe not so bad. I wonder how py3.7 is in comparison
<mattip> {'Ran': 6231, 'failures': 105, 'errors': 17, 'skipped': 1021}
<cfbolz> so, lots of work
<mattip> maybe something is off, they went from 6231 tests to 11013?
<cfbolz> not impossible
<cfbolz> toooons of work (looking through the failures a bit)
<mattip> :(
<cfbolz> well, to be expected
<mattip> the 3.7 -> 3.8 transition was supposed to be a simple
<mattip> one
<cfbolz> mattip: it's always the details that kill us
ronan__ has joined #pypy
Julian has joined #pypy
ronan has quit [Ping timeout: 264 seconds]
ronan__ is now known as ronan
idnar_ is now known as idnar
Julian has quit [Quit: leaving]
glyph has quit [Quit: End of line.]
glyph has joined #pypy
lritter has joined #pypy
lritter has quit [Ping timeout: 252 seconds]
glyph has quit [Read error: Connection reset by peer]
glyph has joined #pypy
glyph has quit [Client Quit]
smarr has joined #pypy
glyph has joined #pypy
glyph has quit [Client Quit]
glyph has joined #pypy
<smarr> hi, has been a while since I last was here. What's the best way to diff RPython trace files? My memory tells me that I might have used a script in the past to remove/filter out things like pointers, which make a basic diff very noise. Though, since it has been a long time, I don't really know anymore what I did. Any hints welcome
glyph has quit [Read error: Connection reset by peer]
glyph has joined #pypy
<smarr> ah, I think I found it, pyptrace.vim
glyph has quit [Client Quit]
glyph has joined #pypy
glyph has quit [Quit: End of line.]
glyph has joined #pypy