<lesshaste>
ah.. it's not new just there has been a new version and it is new to me :)
egregius313 has joined #pypy
<lesshaste>
LarstiQ, I am guessing the answer is no from your general tone
egregius313 has quit [Ping timeout: 276 seconds]
<cfbolz>
it's not impossible that it could be used. we don't know how fast it is to generate code
<cfbolz>
lesshaste: in a sense the problems are going to be similar to the ones of using llvm
black_ant has joined #pypy
<lesshaste>
cfbolz, ah ok. Although I understand time limitations, a full summary of why llvm doesn't work for pypy would be very interesting. I remember that originally the problems were specific flaws in llvm. But those got fixed (I think) and then it turned out the difficulties were more fundamental... I think
<njs>
and if I run it under pypy3 5.8.0 or current nightly, with coverage enabled, then the output is like: "threads_ran is 5 on entry", "threads_ran is 6 on exit", "threads_ran is 5 on entry", "threads_ran is 6 on exit"
<njs>
i.e. one of the modifications to the closed-over variable in one thread, is not seen in another thread, even though it's correctly protected by a lock
<cfbolz>
gah
<cfbolz>
njs: that's scary. bug please?
<njs>
what I'm not sure of is how it usefully report this, since the actual code is this giant pile of stuff and this appears to be extremely fragile
antocuni has joined #pypy
jamescampbell has joined #pypy
lritter has joined #pypy
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
jamescam_ has joined #pypy
jamescampbell has quit [Ping timeout: 240 seconds]
jamescampbell has joined #pypy
jamescam_ has quit [Read error: Connection reset by peer]
realitix has quit [Quit: Leaving]
realitix has joined #pypy
black_ant has quit [Ping timeout: 240 seconds]
marr has joined #pypy
black_ant has joined #pypy
rokujyouhitoma has joined #pypy
rokujyouhitoma has quit [Ping timeout: 255 seconds]
adamholmberg has joined #pypy
jamescampbell has quit [Remote host closed the connection]
<dmalcolm>
IIRC I got the simple cases working, but got lost in the weeds when it came to patching/updating guards
black_ant has quit [Ping timeout: 240 seconds]
<lessshaste>
dmalcolm, thanks!
<lessshaste>
dmalcolm, out of interest.. what are the usability improvements?
<dmalcolm>
lessshaste: general improvements to locations of warnings/errors in the C and C++ frontends is one of the biggies
<lessshaste>
sounds very important!
<lessshaste>
more power to you
<dmalcolm>
lessshaste: tnx. I implemented underlined ranges rather than just points in gcc 6, so basically (much) more of the same
<lessshaste>
very nice
<lessshaste>
I like the clang//
<dmalcolm>
(nods)
<lessshaste>
clang/llvm and gcc are in a usability race :)
marky1991 has joined #pypy
<dmalcolm>
I can spare some cycles to answer qns about libgccjit if anyone's interested in porting things to using it, but I have too much on my plate to be able to devote to being the primary developer for a pypy port
<lessshaste>
I won't have anywhere near enough time/expertise myself but it's a very interesting questiojn
lritter has quit [Remote host closed the connection]
Rhy0lite has joined #pypy
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
<fijal>
dmalcolm: I blame RMS
* fijal
hides
<ronny>
rms created quite a mess because he made sensible monetizion/licensing impossible - and now much opensource happens as the breadcrumbs of enterprises lowering their costs ^^
ArneBab_ has quit [Ping timeout: 246 seconds]
rokujyouhitoma has joined #pypy
ArneBab has joined #pypy
ArneBab has joined #pypy
ArneBab has quit [Changing host]
rokujyouhitoma has quit [Ping timeout: 268 seconds]
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
jamescampbell has quit [Ping timeout: 260 seconds]
<antocuni>
the "self._is_gcc" check fails, so it uses "-R" instead of "-Wl,-R"
jamescam_ has quit [Ping timeout: 246 seconds]
<antocuni>
on cpython, sysconfig.get_config_var('CC') is "x86_64-linux-gnu-gcc", so it works
<antocuni>
on pypy, it's "CC", hence the bug
<antocuni>
now the question is: why on pypy it's "cc"?
rokujyouhitoma has joined #pypy
tilgovi has quit [Ping timeout: 255 seconds]
<antocuni>
ah, it's hard-coded in lib-python/2.7/distutils/sysconfig_pypy.py
jamescampbell has quit [Remote host closed the connection]
rokujyouhitoma has quit [Ping timeout: 268 seconds]
realitix has quit [Ping timeout: 268 seconds]
<nimaje>
well, cc is normaly the default c compiler which should be used (normally override able by the env var CC), but you should also check the expected format for arguments if needed like here
<antocuni>
yes but it's cpython's stdlib code to do that check
<nimaje>
and checking the executable name is not enought :)
<antocuni>
blame cpython :)
lessshaste has quit [Ping timeout: 246 seconds]
pilne has joined #pypy
Ryanar has quit [Quit: Ryanar]
<nimaje>
(why do they not use str.startswith()?)
<antocuni>
to handle things like "x86_64-linux-gnu-gcc", I suppose
lessshaste has joined #pypy
pilne is now known as CodeSnow
<nimaje>
I meant for example ´if sys.platform[:6] == "darwin":´ or ´elif sys.platform[:7] == "freebsd":´
<antocuni>
ah, I don't know
<antocuni>
this is the kind of code which is extremely risky to touch, because a minor change might break things on some obscure platform/distribution/environment/etc
Ryanar has joined #pypy
jacob22 has joined #pypy
jacob22_ has quit [Ping timeout: 246 seconds]
<Alex_Gaynor>
nimaje: If you send a PR to CPython to replace sys.platform[:6] == "darwin" with sys.platform.startswith("darwin"), I can merge it
tilgovi has joined #pypy
tbodt has joined #pypy
tilgovi has quit [Ping timeout: 276 seconds]
adamholmberg has quit [Remote host closed the connection]
asmeurer_ has joined #pypy
tbodt has quit [Client Quit]
adamholmberg has joined #pypy
tbodt has joined #pypy
tbodt has quit [Client Quit]
jamescampbell has joined #pypy
adamholmberg has quit [Ping timeout: 240 seconds]
jamescampbell has quit [Remote host closed the connection]
vkirilichev has quit [Remote host closed the connection]
adamholmberg has quit [Remote host closed the connection]
rokujyouhitoma has quit [Ping timeout: 258 seconds]
adamholmberg has joined #pypy
adamholmberg has quit [Ping timeout: 260 seconds]
jamesaxl has joined #pypy
adamholmberg has joined #pypy
lessshaste has quit [Quit: Leaving]
vkirilichev has joined #pypy
jamescampbell has joined #pypy
<nimaje>
my grep gives 136 lines where startswith could be used ( https://ptpb.pw/s0H0 ), does somebody know if some files in the CPython repo have to support a version where str.startswith() is not supported? and if yes which. if I remember the passphrase for my github ssh key I write a PR
<Alex_Gaynor>
There's no supported version of python missing str.startswith
<xorAxAx>
python is regularly missing str.beginswith, though
vkirilichev has quit [Remote host closed the connection]
vkirilichev has joined #pypy
<nimaje>
well, pep291 says platform and pybench should support python 1.5.2 and for 1.5.2 compatibility you should avoid string methodes
<Alex_Gaynor>
pep291 probably ought to be updated. there's no earthly reason to support python 1.5.2
<marky1991>
lol
rokujyouhitoma has joined #pypy
<xorAxAx>
there is probably at least one bank that still runs python 1.5.2 on their mainframe :)
jimbaker has quit [Ping timeout: 240 seconds]
jimbaker has joined #pypy
_whitelogger has joined #pypy
egregius313 has joined #pypy
jamescampbell has quit [Remote host closed the connection]
egregius313 has quit [Ping timeout: 255 seconds]
jamescampbell has joined #pypy
jamescampbell has quit [Remote host closed the connection]
jamescampbell has joined #pypy
rokujyouhitoma has joined #pypy
tbodt has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
tbodt has joined #pypy
rokujyouhitoma has quit [Ping timeout: 260 seconds]