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
speeder39_ has joined #pypy
dansan_ has joined #pypy
dansan has quit [Remote host closed the connection]
jcea has quit [Quit: jcea]
dansan_ is now known as dansan
<mattip> it turns out buildbot lock(counting) is not consulted when queueing a "scheduled" build,
<mattip> so aarch64 and win32 builds are still running in parallel
<mattip> we seem to be running buildbot 0.8.8 released Aug 2013, the current version is 2.3.1 released May 2019
<mattip> the issue was fixed in buildbot 1.17, http://docs.buildbot.net/latest/relnotes/index.html#id19
<mattip> not clear whether it would be easier to
<mattip> 1) reschedule our builds to ensure they do not overlap
<mattip> 2) hack at the current code to add lock handling to prevent overlap
<mattip> 3) update buildbot to 2.3
speeder39_ has quit [Quit: Connection closed for inactivity]
<mattip> arigato: aarch64 translation segfaulted the pypy interpreter with a NotImplementedError
<mattip> [ARM64/regalloc] instance_ptr_ne not implemented
<mattip> don't know if the interpreter reached some impossible state again or a real problem
<kenaan> mattip buildbot e388d803194c /bot2/pypybuildbot/master.py: change scheduling to reduce overlap on win32, aarch64
<mattip> 1) seems the easiest for now
<mattip> from a short discussion on #buildbot, it seems upgrading from 0.8.8 changes the db model
<mattip> all logs are now stored in the DB, so continuing to use sqlite3 is less of an option and we would have to use a real DB
dddddd has quit [Remote host closed the connection]
<arigato> mattip: I think overlap on aarch64 is not a problem: the machine has got more than enough memory
<arigato> mattip: ...and I think I'm not understanding your point actually
<arigato> mattip: also, I may be missing the point about the "real DB": surely sqlite3 is completely enough to store a few MBs or dozen of MBs of text per day?
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
agronholm has quit [Ping timeout: 248 seconds]
<mattip> arigato: for some reason translation is randomly failing
<mattip> my point about the DB is that someone who did the 0.8 -> 0.9 transition noted that sqlite3 became a bottleneck for them
<arigato> OK
<arigato> translation should be fixed
<kenaan> arigo default 7200528f8f4d /rpython/jit/backend/aarch64/regalloc.py: fix (uh, why support instance_ptr_eq but miss instance_ptr_ne??)
<mattip> py3.6 fails differently
<mattip> AssertionError: live GC var around v4800 = gc_thread_run()!
<arigato> ...bah, one more
<mattip> (and thanks for fixing that one)
<kenaan> arigo default df974ce2019b /rpython/jit/backend/aarch64/opassembler.py: bah, one more
<arigato> the py3.6 failure is going to be harder, because I think that translation works when running it by hand
<arigato> but I can try again
<mattip> grasping at straws - could translation with --batch change anything of importance?
<arigato> no, but randomness can
lritter has joined #pypy
<arigato> also actually maybe I tried with translation running on cpython, not pypy
<mattip> I tried with exactly the buildbot pypy which succeeded, but I guess if we run it enough times it should crash on one of them
<arigato> :-/
<mattip> and, fwiw, the py3.6 buildbot run is not running in parallel with anything else, which strengthens your theory that it is not due to process overlap
<mattip> (weakens my theory that it is due to overlap)
speeder39_ has joined #pypy
<kenaan> mattip buildbot ba7a21c88b5a /bot2/pypybuildbot/master.py: Backed out changeset: e388d803194c
<mattip> on a more positive note, default is green on linux64 :)
<arigato> :-)
<mattip> cfbolz: python improved its syntax error for "print 'abc'" on python3,
<mattip> it raises SyntaxError: Missing parentheses in call to 'print'. Did you mean print('abc')?
<mattip> pypy3.6 still only raises SyntaxError: Missing parentheses in call to 'print'
<mattip> cfbolz: is that something you could possibly look at?
<cfbolz> mattip: yes, I should be able to, after this week (we're moving Tuesday!)
<mattip> thanks, good luck with the move
<cfbolz> Thanks, will need it (why is there so much stuff?!)
_whitelogger has joined #pypy
<mattip> ronan: we have alot of code in lib-python/conftest.py, but AFAIK we still do not support running untranslated lib-python tests
<mattip> is there a reason we can't simplify the lib-python test runner to just call <target> -mpytest lib-python ?
<kenaan> mattip py3.6 2e89e7ae99c0 /: test, fix error message for CPython3 compatibility
<kenaan> mattip py3.6 283ba49cc4e4 /pypy/module/__pypy__/test/test_signal.py: fix test (do not override baseclass setup_class, remove redundant skip clause)
<kenaan> mattip py3.6 a8ced5af1cae /lib-python/3/test/_test_multiprocessing.py: add gc.collect to make tests pass skip test that uses gc.threshold
antocuni has joined #pypy
antocuni has quit [Ping timeout: 272 seconds]
<cfbolz> haha, that print warning about missing parenthesis is a really weird hack:
<cfbolz> it's purely done in the implementation of the SyntaxError class, as opposed to doing it in the parser, where all the correct info would be available
ronan has quit [Quit: Ex-Chat]
ronan has joined #pypy
<cfbolz> mattip: I am confused, pypy3 does at least some part of the print error correctly
<cfbolz> mattip: which test did you see that still fails?
<ronan> mattip: I don't really know what that conftest does, or tries to do, TBH. But IIRC, running the CPython test suite under pytest doesn't work fully
<cfbolz> mattip: I assume test_exception.py?
agronholm has joined #pypy
<mattip> cfbolz: yes, also test_print.py
<cfbolz> ok
<kenaan> cfbolz py3.6 ba0a52e09be0 /pypy/: ImportError exposes a .msg now
<kenaan> cfbolz py3.6 6d08856e48f3 /pypy/module/exceptions/: adopt changes in BaseException.__repr__
speeder39_ has quit [Quit: Connection closed for inactivity]
antocuni has joined #pypy
<mattip> ronan: cpython runs TESTRUNNER in the Makefile with various arguments, where TESTRUNNER is `./python ./Tools/scripts/run_tests.py`
dddddd has joined #pypy
tsaka_ has joined #pypy
nphg1 has joined #pypy
nphg1 has left #pypy [#pypy]
jcea has joined #pypy
antocuni has quit [Ping timeout: 248 seconds]
lritter has quit [Ping timeout: 258 seconds]
tsaka_ has quit [Ping timeout: 248 seconds]
tsaka_ has joined #pypy
marvin_ has quit [Remote host closed the connection]
marvin has joined #pypy
marvin is now known as Guest68312
speeder39_ has joined #pypy
Guest68312 has quit [Remote host closed the connection]
marvin__ has joined #pypy
BPL has joined #pypy
BPL has quit [Read error: Connection reset by peer]
Taggnostr3 has joined #pypy
antocuni has joined #pypy
Taggnostr has quit [Ping timeout: 250 seconds]
antocuni has quit [Ping timeout: 268 seconds]
speeder39_ has quit [Quit: Connection closed for inactivity]
<mattip> arigato: I see a translation of py3.6 succeeded on the screen on arm64-2
<mattip> I started a translation of default (including the changes in 7200528f8f4d) using cpython2
<mattip> assuming it succeeds, I will hopefully set that up as the pypy-for-translation before tonite's runs
jcea1 has joined #pypy
jcea has quit [Read error: Connection reset by peer]
jcea1 is now known as jcea
antocuni has joined #pypy
speeder39_ has joined #pypy
antocuni has quit [Ping timeout: 245 seconds]