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
<antocuni>
every time you run a cpyext test, it prints a number of lines like this:
<antocuni>
missing slot '__setattr__'/'tp_setattr', discovered on 'object'
dddddd has joined #pypy
<antocuni>
what is their purpose? Are they "real" warnings in the sense that they point to things which are not implemented yet? Or it's a leftover of some older dev stuff which we could just kill?
<antocuni>
it seems that slotdefs is populated from code which is copied&pasted from CPython, so I guess that the missing slots are meant to be missing?
jcea has joined #pypy
jcea has quit [Ping timeout: 248 seconds]
lritter has joined #pypy
jcea has joined #pypy
<Dejan>
antocuni, could you please build 373 versions of your wheels? :)
<Dejan>
hopefully latest numpy too
ssbr` has quit [Remote host closed the connection]
jcea has quit [Remote host closed the connection]
<Dejan>
weird error when trying to install blist (latest build of PyPy 3.6)
<Dejan>
blist/_blist.c:166:17: error: unknown type name ‘PY_UINT64_T’
<Dejan>
166 | PY_UINT64_T k_uint64;
fryguybob has joined #pypy
<Dejan>
IMHO it is a bug
<Dejan>
CPython 3.6 defines it
<cfbolz>
sounds like it, yes
<Dejan>
there are several definitions that are missing in pyport.h
<Dejan>
cfbolz, could you please add the following to it:
<Dejan>
#define PY_UINT32_T uint32_t
<Dejan>
#define PY_INT32_T int32_t
<Dejan>
#define PY_UINT64_T uint64_t
<Dejan>
#define PY_INT64_T int64_t
<cfbolz>
Dejan: please open a pull request, I know nothing about this part of pypy, sorry :-(
<Dejan>
I would but I really can't bother with mercurial
<Dejan>
and bitbucket
<cfbolz>
life is hard
<Dejan>
I just logged on bitbucket... and got greeting msg from them they are removing Mercurial support this year
<cfbolz>
yes, we know. we are still evaluating options
<Dejan>
fossil-scm?
<Dejan>
J/K
<Dejan>
1838 branches...
<Dejan>
Jesus...
<cfbolz>
:-)
<Dejan>
Here at BMLL where I work, when we get over 30 branches we have an alarm :)
<Dejan>
sure PyPy is a different story
<antocuni>
Dejan: pypy uses named branches, so even those who were merged are still counted in the history
<Dejan>
antocuni, did you see my msg above?
<antocuni>
which one?
<Dejan>
Could you also add blist there too? :D
<Dejan>
<Dejan> antocuni, could you please build 373 versions of your wheels? :)
<Dejan>
<Dejan> hopefully latest numpy too
<antocuni>
ah no, I missed those messages
<antocuni>
thanks for the reminder, I'll add it. But again, PR are welcome :)
<Dejan>
I can do that
<Dejan>
(as your stuff are on git)
<antocuni>
thanks
<Dejan>
in pypy-wheels ?
<antocuni>
yes
<cfbolz>
it's admittedly also true that we have a ton of open kind of abandoned branches
kenaan_ has joined #pypy
jcea has joined #pypy
Rhy0lite has joined #pypy
adamholmberg has joined #pypy
<Dejan>
antocuni, since it states "for two latest PyPy releases", should I remove 7.1.1 and add 7.2.0 ?
<antocuni>
yes, I think so
<Dejan>
also there are some failed builds you should have a look when you have time :)
<antocuni>
yes, I know... builds keep failing from time to time :(
ronan__ has joined #pypy
Smigwell has joined #pypy
Smigwell has left #pypy [#pypy]
<ronan__>
antocuni: there's a few slots that are actually missing on object, but they're listed explicitly in the code
<ronan__>
IIRC, mattip found the warnings useful sometimes, for other types
<ronan__>
but I'm not sure they're worth the hassle
<ronan__>
Dejan: has blist ever worked on pypy? These defines have been missing forever
<Dejan>
ronan__, this is the first time I am trying
<Dejan>
ronan__, could you please add those to pyport.h please as for me to make the PR will take ages
<Dejan>
pull-requests or merge-requests, whatever they prefer as a name, should be extended so one can simply sand a patch file
<Dejan>
s/sand/send/
<ronan__>
Dejan: yes, I'm looking at it now
<Dejan>
ronan__, thanks
<ronan__>
though I may not manage to do it today
jvesely has joined #pypy
<Dejan>
sure, no worries
jcea has quit [Remote host closed the connection]
<ronan__>
Dejan: should be fixed on pypy2 but my train is almost in Paddington... I'll try to merge to py3.6 tonight so you should be able to try a nightly tomorrow
ronan__ has quit [Remote host closed the connection]
jcea has joined #pypy
<Dejan>
IMHO our binary 3.x packages hould contain a symbolic link pypy -> pypy3
<Dejan>
unless you want it to be possible to untar both 2.7 and 3.x in the same dir
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
adamholmberg has quit [Read error: Connection reset by peer]
adamholmberg has joined #pypy
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
micisuta has quit [Ping timeout: 240 seconds]
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
jcea has quit [Remote host closed the connection]
oberstet has quit [Quit: Leaving]
<Dejan>
ronan, we should go for a pint some day when you are in London
<Dejan>
:)
<Dejan>
antocuni:
<Dejan>
blist/_blist.c:166:17: error: unknown type name ‘PY_UINT64_T’
<Dejan>
PY_UINT64_T k_uint64;
<Dejan>
i got the same error, as expected
<Dejan>
but at least I got the other wheels
<Dejan>
well, i only tried numpy actually
micisuta has joined #pypy
ronan__ has joined #pypy
<Dejan>
time to go home
<Dejan>
:)
jcea has joined #pypy
ronan__ has quit [Remote host closed the connection]
ronan__ has joined #pypy
jcea has quit [Remote host closed the connection]
ronan__ has quit [Remote host closed the connection]
jcea has joined #pypy
ronan__ has joined #pypy
<ronan__>
Dejan: ping
jvesely has quit [Remote host closed the connection]
jvesely has joined #pypy
ronan__ has quit [Quit: Leaving]
ssbr` has joined #pypy
Rhy0lite has quit [Quit: Leaving]
ronan__ has joined #pypy
ronan__ has quit [Remote host closed the connection]
jvesely has quit [Quit: jvesely]
jvesely has joined #pypy
ronan__ has joined #pypy
xcm has quit [Killed (verne.freenode.net (Nickname regained by services))]
xcm has joined #pypy
TheNewbie has joined #pypy
micisuta1 has joined #pypy
<ronan__>
Dejan: it should be fixed now but blist isn't going to work on pypy3 anyway, because it uses CPython internals
<ronan__>
and btw blist is unmaintained and broken on CPython 3.7+
micisuta has quit [Ping timeout: 258 seconds]
micisuta1 is now known as micisuta
ronan__ has quit [Ping timeout: 265 seconds]
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
xcm has quit [Read error: Connection reset by peer]