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
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
lritter has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
inhahe has quit [Read error: Connection reset by peer]
inhahe has joined #pypy
inhahe has quit [Read error: Connection reset by peer]
inhahe has joined #pypy
inhahe has quit []
inhahe has joined #pypy
oberstet has joined #pypy
oberstet has quit [Ping timeout: 240 seconds]
oberstet has joined #pypy
tsaka__ has joined #pypy
dddddd has joined #pypy
tsaka__ has quit [Ping timeout: 245 seconds]
<mattip>
the aarch64 problem with template parsing is strange (issue 3104). It does recursive parsing, adding more pieces of text to a list at each level
<mattip>
and in the end returns ''.join(lst)
<mattip>
when run inside the setup.py, something seems to overflow so some of the lst is missing
<mattip>
but when I run stand-alone, it all works well
<arigato>
mattip: maybe try to rule out another gcc problem? run a pypy-c that was compiled with "make lldebug0" and see if it (slowly) works or not
<arigato>
or with a gcc < 8, I think
<mattip>
maybe. But then why would it only happen inside the setup.py/distutils call (--jit off)?
<arigato>
I have no clue
<arigato>
I can ask the same question about most explanations, though
<arigato>
if you want, I can try to dig inside gdb. From a reproducer it should "only" be a matter of enough digging
jvesely has joined #pypy
<mattip>
I am trying to get it down to a reproducer, it seems to need too much of the surrounding code
<mattip>
which is also suspicious
<arigato>
yes, it's strange (in "--jit off" mode at least)
<arigato>
maybe something with the GC? there's a malloc somewhere that one in a while follows the slow path "nursery full, need to trigger collection", and the bug is in this slow path
<arigato>
s/one in a while/once in a while
Taggnostr has quit [Remote host closed the connection]
<mattip>
ok, I will try adding some gc.collect(). Also trying to move the global re.compile expressions to local, maybe something is corrupting them
<arigato>
try to run with an env var
<arigato>
PYPY_GC_something
<mattip>
+1
Taggnostr has joined #pypy
<arigato>
PYPY_GC_NURSERY=1
<arigato>
(very slow)
* mattip
bbl
tsaka__ has joined #pypy
dustinm- has joined #pypy
dustinm has quit [Ping timeout: 246 seconds]
<ajlawrence>
What is the best way to handle unions with rffi?
tsaka__ has quit [Ping timeout: 276 seconds]
jacob22_ has joined #pypy
jacob22 has quit [Read error: Connection reset by peer]
<arigato>
ajlawrence: good question
<arigato>
it's a hack, but you can simply declare them as structs
<ajlawrence>
arigato: I will give it a try and see how it goes.
tsaka__ has joined #pypy
xorAxAx_ has joined #pypy
xorAxAx_ has left #pypy [#pypy]
<mattip>
with PYPY_GC_NURSERY=1 the templating produced the proper output
xorAxAx has joined #pypy
* xorAxAx
is currently playing with a simple LLVM-IR to rpython transpiler
marvin_ has quit [Remote host closed the connection]
marvin_ has joined #pypy
tsaka__ has quit [Ping timeout: 250 seconds]
<ajlawrence>
Is it possible for the translation to become stuck in a loop?
<mattip>
probably not
<mattip>
has it been running for more than an hour on a newish machine?
<ajlawrence>
No. It has just stopped producing the fractal.
<mattip>
at one point something was messed up with the ascii output on windows. I usually do "set PY_COLORS=0"
<mattip>
if so, then the output would slowly scroll out of view, since there is output to the screen but the colors are all black
<kenaan>
mattip default 17694ed47a34 /lib-python/2.7/ensurepip/__init__.py: backport ensurepip from 3.6
jacob22 has joined #pypy
jacob22_ has quit [Read error: Connection reset by peer]
oberstet has quit [Remote host closed the connection]