cfbolz changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | the secret reason for us trying to get PyPy users: to test the JIT well enough that we're somewhat confident about it
<njs>
BTW, I had a thought for something you might want to think about during the sprint: now that cpyext kinda mostly works, it'd be nice if PyPy could sort out the wheel ABI tagging stuff so that it would be possible for projects to start uploading pypy wheels to pypi
LooCfur has quit [Read error: Connection reset by peer]
LooCfur has joined #pypy
jamesaxl has joined #pypy
andrew2nelson has joined #pypy
Remi_M has joined #pypy
<andrew2nelson>
hey chan, I'm looking into patching MiniMarkGC to keep all object flags in their own page (to keep shared memory clean during major collection). If I create an AddressDict member in MiniMarkGC that maps an object's address to its set of flags, does that accomplish this?
asmeurer_ has quit [Quit: asmeurer_]
Remi_M has quit [Ping timeout: 256 seconds]
tayfun26 has joined #pypy
<mjacob>
andrew2nelson: there already is a branch where this was started. let me look for it...
<mjacob>
andrew2nelson: btw, most probably you want to patch IncrementalMiniMarkGC
<mjacob>
andrew2nelson: the branch is called gc-forkfriendly-2
<andrew2nelson>
Thanks! I'll take a look
<andrew2nelson>
Is IncrementalMiniMarkGC currently the default? I see conflicting reports from the docs
<mjacob>
possibly the docs are outdated
<mjacob>
but yes, incminimark is the default
<andrew2nelson>
good to know
<mjacob>
if you point me at the outdated places in the docs, i'll fix them
mcyprian has joined #pypy
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<kenaan>
Raemi guard-compatible aab085ce1480 /rpython/jit/metainterp/: (arigato, remi) cleanup test and make the bridge not contain a call_pure By preventing to share descrs wi...
<fijal>
andrew2nelson: feel free to complain about the conflicting reports - we will rectify that
amaury__ has quit [Quit: Konversation terminated!]
amaury__ has joined #pypy
<kenaan>
Raemi guard-compatible f55580d6252b /rpython/jit/metainterp/test/test_compatible.py: (arigato, remi) it seems to us that the current output of the test is correct and better
<arigato>
cfbolz: happiness, it seems
<cfbolz>
arigato: ah?
<arigato>
yes, just commenting about aab085ce1480
<arigato>
waiting for a translation now
<cfbolz>
Awesome
<cfbolz>
arigato: thanks a lot :-)
<arigato>
yesterday we got "pypy richards.py" to run with 3 less branches than on default
<arigato>
where "we" == "remi"
<cfbolz>
arigato: a good benchmark to try is to run metainterp's test_recursive.py on PyPy
Remi_M has quit [Ping timeout: 276 seconds]
antocuni has joined #pypy
marr has joined #pypy
mcyprian has joined #pypy
amaury__ has quit [Ping timeout: 240 seconds]
inad922 has joined #pypy
<fijal>
arigato: FYI sketchup tweeted about us
mcyprian has quit [Ping timeout: 256 seconds]
mcyprian has joined #pypy
inad922 has quit [Ping timeout: 260 seconds]
beystef has joined #pypy
inad922 has joined #pypy
Taggnostr2 has joined #pypy
Taggnostr has quit [Ping timeout: 264 seconds]
Taggnostr has joined #pypy
Taggnostr2 has quit [Ping timeout: 276 seconds]
oberstet has joined #pypy
Taggnostr has quit [Ping timeout: 240 seconds]
Taggnostr has joined #pypy
Taggnostr has quit [Ping timeout: 260 seconds]
Taggnostr has joined #pypy
lritter has joined #pypy
Taggnostr2 has joined #pypy
Taggnostr has quit [Ping timeout: 248 seconds]
Taggnostr has joined #pypy
Taggnostr2 has quit [Ping timeout: 276 seconds]
tormoz has quit [Read error: Connection reset by peer]
geov has joined #pypy
Taggnostr2 has joined #pypy
Taggnostr has quit [Ping timeout: 264 seconds]
inad922 has quit [Ping timeout: 264 seconds]
Taggnostr has joined #pypy
Taggnostr2 has quit [Ping timeout: 256 seconds]
Taggnostr has quit [Ping timeout: 260 seconds]
Taggnostr has joined #pypy
inad922 has joined #pypy
Remi_M has joined #pypy
mcyprian has quit [Ping timeout: 276 seconds]
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
illume has joined #pypy
dddddd has joined #pypy
<arigato>
fijal: cool
mcyprian has joined #pypy
<mjacob>
fijal: any progress on the "sprint in poland" front?
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<kenaan>
mjacob reverse-debugger-updated 0365344dd2f6 /rpython/: Fix replaying by adding back hash field to GC header.
mcyprian has quit [Quit: Leaving.]
demonimin has quit [Ping timeout: 264 seconds]
<kenaan>
antocuni cpyext-fast-typecheck b8745d3e1669 /pypy/module/cpyext/test/test_typeobject.py: add a passing test
<kenaan>
antocuni cpyext-fast-typecheck a0cad484bfaf /pypy/module/cpyext/methodobject.py: move this logic inside W_PyCWrapperObject.descr_call, for consistency with the other classes, and ...
<kenaan>
antocuni cpyext-fast-typecheck ba7b2efa9585 /pypy/module/cpyext/: WIP: modify the W_PyCWrapperObject call logic: instead of using a very generic logic which always ...
mcyprian has joined #pypy
illume has joined #pypy
mcyprian has quit [Quit: Leaving.]
mcyprian has joined #pypy
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
mcyprian has quit [Ping timeout: 240 seconds]
mcyprian has joined #pypy
<kenaan>
antocuni cpyext-fast-typecheck a3b4293a2b1b /pypy/module/cpyext/test/test_cpyext.py: (anto, ronan): fix this test, because gcc (correctly :)) complains that the statement had no effect
Rhy0lite has joined #pypy
<kenaan>
mjacob reverse-debugger-updated 899d967fe147 /rpython/memory/gctransform/boehm.py: Change 'hash' field to be there only for revdb.
amaury__ has joined #pypy
jcea has joined #pypy
<kenaan>
arigo vtune 2b28145a01f9 /rpython/jit/backend/x86/vtune.py: updates for 2018
amaury__ has quit [Quit: Konversation terminated!]
amaury has joined #pypy
AndrewBC has joined #pypy
<kenaan>
antocuni cpyext-fast-typecheck 5ef7b9a78fc4 /pypy/module/cpyext/: WIP, refactor the previous commit: since wrap_binaryfunc has a different signature, we can no long...
mcyprian has quit [Ping timeout: 240 seconds]
<cfbolz>
arigato: vtune, advanced
<arigato>
yes, but it's the only way we found to get per-instruction timings in jit code
<kenaan>
rlamy py3tests e703f1b77506 /pypy/conftest.py: Collect apptest_*.py files also when untranslated
<kenaan>
rlamy py3tests aaa43023ad0d /pypy/: Add option to enable app-level assert rewriting; turn it off by default
<cfbolz>
arigato: any interesting insights?
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
raynold has quit [Quit: Connection closed for inactivity]
mcyprian has joined #pypy
amaury has quit [Quit: Konversation terminated!]
jamesaxl has quit [Read error: Connection reset by peer]
jamesaxl has joined #pypy
amaury has joined #pypy
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
marky1991 has joined #pypy
jcea has quit [Remote host closed the connection]
marky1991 has quit [Remote host closed the connection]
tayfun26 has quit [Remote host closed the connection]
<kenaan>
antocuni cpyext-fast-typecheck 430955eada2e /pypy/module/cpyext/: WIP: refactor again: there is no point in having tons of wrappers and a different class for each w...
<kenaan>
antocuni cpyext-fast-typecheck 624c476bc7cd /pypy/module/cpyext/: test and fix typo
<kenaan>
Raemi guard-compatible ad7291e0a4d3 /rpython/jit/backend/x86/: cleanup and fix for passing non-code data to vtune
<kenaan>
antocuni cpyext-fast-typecheck 8b23d59cea09 /pypy/module/cpyext/: make sure to raise TypeError if you pass keywords to a slot which does not expect them
<kenaan>
rlamy py3tests bc7db4a511cc /pypy/tool/pytest/apptest2.py: Implement AppTestFunction.reportinfo() to get correct failure reports
<lesshaste>
fijal, "Everyone is talking about it, looking at it, investigating etc." I would write "Everyone is talking about, looking at and investigating it."
<lesshaste>
typo: insterested
<lesshaste>
typo: is an incredibly improvement
<lesshaste>
typo: we draw lines with mouse
<lesshaste>
typo: Doing visualization that look good
<lesshaste>
typo: one starts feeling taking the helmet
<arigato>
fijal: a more "interesting" typo: "Oddysey" => "Odyssey"
mcyprian has quit [Quit: Leaving.]
jamesaxl has quit [Read error: Connection reset by peer]
<cfbolz>
:-)
<cfbolz>
though it was probably pretty odd too
jamesaxl has joined #pypy
<illume>
matplotlib (release and master) don't seem to work on mac. The dreaded undeclared 'PyOS_InputHook'.
<illume>
I will try to make a stub PyOS_InputHook.
<arigato>
fijal: otherwise, looks good to me (even if the last section is probably too optimistic about 1 release every 3 weeks)
<fijal>
I'll change it to two months?
asmeurer__ has joined #pypy
<kenaan>
antocuni cpyext-fast-typecheck 0e0dc6fb5d4c /pypy/module/cpyext/slotdefs.py: convert wrap_objobjproc to the new style
<kenaan>
antocuni cpyext-fast-typecheck dc0945ef4c26 /pypy/module/cpyext/slotdefs.py: rewrite wrap_unaryfunc in the new style
<kenaan>
antocuni cpyext-fast-typecheck edf304aa6652 /pypy/module/cpyext/: port wrap_call to the new style, reusing most of the logic we use also W_PyCFunction.call_keywords
<kenaan>
antocuni cpyext-fast-typecheck de9fbc68ef97 /pypy/module/cpyext/slotdefs.py: go back to the previous naming convention, with a comment explaining why
<kenaan>
antocuni cpyext-fast-typecheck 259e93d2de92 /pypy/module/cpyext/slotdefs.py: port wrap_init to the new style
<kenaan>
antocuni cpyext-fast-typecheck 0e2721692722 /pypy/module/cpyext/: kill wrapper_func_kwds, since it was used only by wrap_init and wrap_call, which are now new-style...
<kenaan>
antocuni cpyext-fast-typecheck 1525006c89ef /pypy/module/cpyext/slotdefs.py: port wrap_setattr and wrap_delattr to the new style
<cfbolz>
Remi_M: did you try to run the new executable on test_recursive.py yet? because for me it segfaults :-(
<andrew2nelson>
mjacob: I think the documentation I was looking at was outdated. Switching to the documentation in the /docs folder looks like it says incminimark across the board
<andrew2nelson>
Is there a way to do a typecheck before translation? My unit tests pass in a minute or two but it looks like they don't catch type errors that show up during translation
andrew2nelson has quit [Ping timeout: 260 seconds]
<kenaan>
antocuni cpyext-fast-typecheck 2e882947bdd3 /pypy/module/cpyext/slotdefs.py: port more wrappers, starting to fix test_arraymodule.py
<kenaan>
antocuni cpyext-fast-typecheck 567c1b64d3ec /pypy/module/cpyext/slotdefs.py: port some of the buffer-related slots, enough to pass the array tests
<kenaan>
antocuni cpyext-fast-typecheck 5baee26aad6e /pypy/module/cpyext/slotdefs.py: port wrap_getattro
jamesaxl has quit [Quit: WeeChat 2.0.1]
getxsick_ has joined #pypy
inhahe_ has quit []
inhahe has joined #pypy
getxsick_ has quit [Quit: leaving]
amaury has joined #pypy
<kenaan>
antocuni cpyext-fast-typecheck 564712927aef /pypy/module/cpyext/: add a test for sq_ass_slice and port the corresponding wrapper to the new style
<kenaan>
antocuni cpyext-fast-typecheck 150b6a41b9f2 /pypy/module/cpyext/: test for sq_ass_item and fix
<kenaan>
antocuni cpyext-fast-typecheck 0db3ba6fe311 /pypy/module/cpyext/: test and fix for wrap_sq_delitem
<kenaan>
antocuni cpyext-fast-typecheck f51b4cef71b0 /pypy/module/cpyext/slotdefs.py: port the last slot wrapper
<kenaan>
antocuni cpyext-fast-typecheck b4908ae52515 /pypy/module/cpyext/: finally kill all the remainings of the old wrappers :)
<kenaan>
antocuni cpyext-fast-typecheck cb1e9314536a /pypy/module/cpyext/methodobject.py: try to see if this improve the performance
hniksic has quit [Quit: Page closed]
illume has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]