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
ronan has quit [Ping timeout: 265 seconds]
marky1991 has quit [Read error: Connection reset by peer]
marky1991 has joined #pypy
ronan has joined #pypy
marky1991_2 has joined #pypy
marky1991_2 has quit [Remote host closed the connection]
marky1991_2 has joined #pypy
marky1991 has quit [Ping timeout: 265 seconds]
marky1991_2 has quit [Ping timeout: 240 seconds]
CrazyPython has joined #pypy
marky1991_2 has joined #pypy
CrazyPython has quit [Read error: Connection reset by peer]
xcm has quit [Remote host closed the connection]
xcm has joined #pypy
marky1991_2 has quit [Ping timeout: 240 seconds]
ronan has quit [Ping timeout: 276 seconds]
asmeurer__ has quit [Quit: asmeurer__]
CrazyPython has joined #pypy
CrazyPython has quit [Read error: Connection reset by peer]
lritter has quit [Ping timeout: 240 seconds]
lritter has joined #pypy
jvesely has joined #pypy
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
jcea has quit [Quit: jcea]
mattip_ has joined #pypy
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4960 [mattip: force build, py3.6]
_whitelogger has joined #pypy
tsaka_ has quit [Ping timeout: 276 seconds]
mattip_ has quit [Ping timeout: 265 seconds]
hotpot33 has quit [Read error: Connection reset by peer]
hotpot33 has joined #pypy
dddddd has quit [Remote host closed the connection]
_whitelogger has joined #pypy
tsaka_ has joined #pypy
_whitelogger has joined #pypy
jvesely has quit [Quit: jvesely]
tsaka_ has quit [Ping timeout: 276 seconds]
tsaka_ has joined #pypy
_whitelogger has joined #pypy
_whitelogger has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4960 [mattip: force build, py3.6]
tsaka_ has quit [Ping timeout: 276 seconds]
ekaologik has joined #pypy
tsaka_ has joined #pypy
<kenaan> arigo default 41990e65c077 /pypy/module/_cffi_backend/newtype.py: oops, aarch64 fix
ronan has joined #pypy
ronan has quit [Ping timeout: 240 seconds]
<kenaan> arigo int-test-is-zero 092643ba66ac /rpython/jit/: Generate more compact code for bit tests using the TEST instruction: i1 = int_and(i0, 7); i2 = int_i...
ronan has joined #pypy
<kenaan> arigo int-test-is-zero 56ea3adf27c4 /rpython/jit/: Make the usage of INT_TEST_xxx conditional on the backend
ronan has quit [Ping timeout: 240 seconds]
<kenaan> arigo int-test-is-zero abbcc2ad6b7f /pypy/module/pypyjit/test_pypy_c/test_misc.py: Add a test_pypy_c test for int_test_is_true
<kenaan> arigo int-test-is-zero f2afd59f9872 /pypy/module/pypyjit/test_pypy_c/test_misc.py: merge heads
ronan has joined #pypy
<kenaan> arigo int-test-is-zero bbc157fc2398 /pypy/module/pypyjit/test_pypy_c/test_misc.py: skip this test on non-x86 platforms
<kenaan> arigo int-test-is-zero d731a41aae59 /: Close branch int-test-is-zero
<kenaan> arigo default 18443d3a74d5 /: Merged in int-test-is-zero (pull request #684) Use the TEST assembler instruction on x86
<kenaan> rlamy hpy e2a687fd8ab3 /pypy/module/cpyext/: Fix cparser issue with struct types appearing in function arguments
CrazyPython has joined #pypy
<kenaan> arigo default 8fc092d343a0 /pypy/module/pypyjit/test_pypy_c/test_micronumpy.py: Fix some failures of test_pypy_c on aarch64
<cfbolz> arigato: cool,
<cfbolz> Did you see a lot of int_and?
lritter has quit [Ping timeout: 265 seconds]
ronan has quit [Ping timeout: 276 seconds]
<arigato> cfbolz: unclear, but over the years it's a pattern I saw from time to time
<arigato> today I saw it in aarch64 to check if a pointer is correctly aligned, but of course I didn't implement that for aarch64 anyway, only for x86 (where it removes more instructions)
<arigato> it shows up on app-level code like "if i % 4:"
<arigato> ...bah
<arigato> in such app-level code, the value "i % 4" appears in failargs so it needs to be computed anyway
<arigato> ...and bah again, in "if (i % 4) == 0:" we get a "int_eq(.., 0)" instead of "int_is_true(..)"
ronan has joined #pypy
<arigato> ok so it looks like a bit of a useless idea
inhahe_ has joined #pypy
CrazyPython has quit [Ping timeout: 245 seconds]
inhahe has quit [Ping timeout: 252 seconds]
CrazyPython has joined #pypy
ronan has quit [Ping timeout: 265 seconds]
ronan has joined #pypy
<cfbolz> arigato: but "if a & 4: " works?
dddddd has joined #pypy
ronan has quit [Ping timeout: 240 seconds]
<arigato> no
<arigato> it does stupidly "i1 = int_and(a, 4); i2 = int_test_is_zero(a, 4); guard(i2)[i1]"
<arigato> not really worse than before but not better either
CrazyPython has quit [Read error: Connection reset by peer]
<arigato> maybe I should revert and try to do a different hack to detect that after computing "i1 = int_and(a, 4)" I can directly check the cpu flags---iow, the "int_is_true" is pointless
ronan has joined #pypy
rubdos has quit [Ping timeout: 250 seconds]
<cfbolz> arigato: yes, sounds like a good idea
<cfbolz> We need tree pattern matching in the x86 backend ;-)
<arigato> :-)
rubdos has joined #pypy
ronan has quit [Ping timeout: 240 seconds]
jvesely has joined #pypy
ronan has joined #pypy
<kenaan> rlamy hpy 01381639f0d0 /pypy/module/cpyext/test/test_cparser.py: Add a cparser test
<kenaan> rlamy hpy 54d6b14f3362 /pypy/module/hpy_universal/: Begin implementing hpy_universal.load()
ronan has quit [Ping timeout: 265 seconds]
<kenaan> arigo hpy 30944126f7e1 /pypy/module/cpyext/api.py: (ronan, arigo) Don't call PyErr_Occurred() when its result is not needed
<kenaan> arigo hpy 5d81fcd424a5 /pypy/module/hpy_universal/: (ronan, arigo) in-progress: start to make C stuff and link it to RPython stuff (for now, that's mostly the HPyContext)
ronan has joined #pypy
ronan has quit [Ping timeout: 276 seconds]
CrazyPython has joined #pypy
ronan has joined #pypy
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
<kenaan> arigo hpy b3357dcf16a7 /pypy/module/hpy_universal/: (ronan, arigo, antocuni just arrived now) Yay, we can now import the empty module
<kenaan> arigo hpy 72227d40aa42 /pypy/module/hpy_universal/test/test_basic.py: Put the next test to pass
ronan__ has joined #pypy
ronan__ has quit [Remote host closed the connection]
ronan__ has joined #pypy
ronan has quit [Ping timeout: 240 seconds]
jcea has joined #pypy
squeaky_pl has joined #pypy
<squeaky_pl> I wonder if somebody studied impact of `-fno-semantic-interposition` on PyPy in interpreter mode. I can imagine it would have not matter mostly with jitted code. https://twitter.com/VictorStinner/status/1195476757048102913
<squeaky_pl> or it doesnt matter at all because all the symbols are already in one .so?
<cfbolz> squeaky_pl: haha, I was wondering the same
<cfbolz> Worth a try!
<cfbolz> I am very sure that nobody sensibly uses LD_PRELOAD on rpython code
CrazyPython has quit [Read error: Connection reset by peer]
<squeaky_pl> https://stackoverflow.com/questions/35745543/new-option-in-gcc-5-3-fno-semantic-interposition does rpython do __attribute__ stuff on the C code it generates?
<cfbolz> squeaky_pl: I don't remember, I think we don't export most functions
<cfbolz> That would mean the 'it's much less useful' case applies
ronan has joined #pypy
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
ronan has quit [Remote host closed the connection]
ronan__ has quit [Ping timeout: 240 seconds]
ronan has joined #pypy
<kenaan> rlamy hpy b0b337fc4202 /pypy/module/hpy_universal/: Replace hpy_universal.load() with hpy_universal.load_from_spec()
<kenaan> arigo hpy 3612658138c2 /pypy/module/hpy_universal/: (antocuni, ronan, arigo) Progress towards, but not yet, reaching test_noop_function
<bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7816 [ronan: force build, hpy]
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
ronan__ has joined #pypy
ronan has quit [Ping timeout: 276 seconds]
CrazyPython has joined #pypy
<squeaky_pl> I wanted to report a bug to PyPy but I can't seem to login to bitbucket (get 500): Here is a reproducer: https://gist.github.com/squeaky-pl/2920bb5f472f70346c57e6a859ecda04
mattip_ has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7816 [ronan: force build, hpy]
ronan has joined #pypy
ronan__ has quit [Ping timeout: 245 seconds]
CrazyPython has quit [Read error: Connection reset by peer]
CrazyPython has joined #pypy
mattip_ has quit [Ping timeout: 276 seconds]
CrazyPython has quit [Read error: Connection reset by peer]
ronan__ has joined #pypy
ronan has quit [Ping timeout: 240 seconds]
tos9 has quit [Read error: Connection reset by peer]
tos9 has joined #pypy
<kenaan> antocuni hpy bacbce0ca621 /pypy/module/hpy_universal/: (antocuni, arigo): implement HPyNone_Get, test_noop_function finally passes
<kenaan> antocuni hpy 949c501dadd7 /pypy/module/hpy_universal/: (antocuni, arigo, ronan): kill the usage of @slot_function from cpyext and use the much simpler @apifunc
<kenaan> arigo hpy f3ead38c0c67 /pypy/module/hpy_universal/: Translation fixes. Try following these specialize.memo() if you can!
<kenaan> antocuni hpy 256add727d58 /pypy/module/hpy_universal/test/test_basic.py: skip this for now
ekaologik has quit [Read error: Connection reset by peer]
<mjacob> squeaky_pl, cfbolz: pypy uses "__attribute__((visibility("hidden")))" on most of it's "internal" functions
<mjacob> so i'd expect that -fno-semantic-interposition wouldn't have a big impact
ronan has joined #pypy
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
ronan__ has quit [Ping timeout: 265 seconds]
ronan__ has joined #pypy
ronan has quit [Ping timeout: 240 seconds]