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
adamholmberg has quit [Remote host closed the connection]
ronan has quit [Remote host closed the connection]
ronan has joined #pypy
* arigato
fixes multibytecodec
<Dejan>
could you also implement PyCodec_Encoder and Decoder ? ;)
<Dejan>
that may render psycopg2 usable on PyPy
<kenaan>
arigo default 2ed84f7866b6 /pypy/module/_multibytecodec/: Fix a corner case in multibytecodec: for stateful codecs, when encoding fails and we use replacement, the replaceme...
<Dejan>
Could you merge your code to 3.6 as that is what we use here...
<Dejan>
getting rid of these problems may make me persuade my colleagues we should move to PyPy 3.6
<cfbolz>
patience
<cfbolz>
Dejan: note that you started a build on py3.7
<Dejan>
yea, i noticed it did not build in a long time, and wanted to test the build
<Dejan>
as i have never done it before
<Dejan>
I assume you will trigger the build once you think it is ready :)
<Dejan>
or it is automated for 3.6?
<cfbolz>
arigato: aaah, @pytest.mark.pypy_only is a py3k thing exclusively
<cfbolz>
I was very confused why it wouldn't fail on default
<cfbolz>
Dejan: for 3.6 it's nightly
dddddd has joined #pypy
<cfbolz>
arigato: I fixed the pytest issue, but need to leave now, could you do the multibytecodecs merge? otherwise I'll try later
<kenaan>
cfbolz py3.6 deaa30ceb571 /pypy/conftest.py: fix issue with @pytest.mark.pypy_only before it would actually crash
<cfbolz>
I am just very confused
<arigato>
cfbolz: I'll do
<arigato>
note that you can say "hg merge 075c10ca9276"
<arigato>
which merges up to and including the test_argument change
<cfbolz>
Right
<cfbolz>
arigato: but the crash had nothing to do with the merge actually
<arigato>
OK
<arigato>
so yes, let me try again the full merge
<kenaan>
arigo default 76b34bb85fe5 /pypy/interpreter/test/test_argument.py: Test really just what we want to, i.e. that the message does not contain the word 'self'
<mattip>
EWDurbin: the DNS is managed by fijal, I am probably the best one to talk to about the website beyond that
<EWDurbin>
cool, mattip two questions on website first then :)
<mattip>
?
<EWDurbin>
1) what's the plan for bitbucket deprecating hg? that's happening in June so we should probably consider it sooner than later
<EWDurbin>
2) the host that's currently serving pypy.org is on some pretty old PSF infra that i'm trying to get us all migrated out of. are there any plans for the site to be anything aside from generated static HTML in the future? just for planning where to migrate it to (which would require DNS updates)
<mattip>
if not we will, like everybody, migrate to github.com
<mattip>
I am involved in the transition and will issue a PR to the chef recipe when we have a solution
<Dejan>
gitlab ftw!
<mattip>
2) The plans for the site are to update it, but to remain with static html
<mattip>
there are some thoughts about eventually moving our blog posts from morepypy.blogspot.com, but even then not clear if we will keep a discus reply section
dddddd has quit [Ping timeout: 260 seconds]
xcm has quit [Remote host closed the connection]
<mattip>
it seems that posting reponses on the blog is going out of style, in favor of mail/twitter/reddit/IRC
<mattip>
</end>
<EWDurbin>
cool thanks mattip! hopefully i can get in touch with fijal about DNS updates soon. as far as i'm aware the only ones the PSF infra needs to be worried about are `pypy.org`, `www.pypy.org`, and `speed.pypy.org`
xcm has joined #pypy
<mattip>
thanks
<mattip>
that seems about right as far as I know. There is buildbot.pypy.org, but that runs on non-PSF hardware managed by fijal.
oberstet has joined #pypy
<mattip>
arigato: thanks for the multibyte fix. It looks so easy when you know the solution
<Dejan>
mattip, that is why i triggered the build - i am eager to try to use psycopg2 with the latest code