cfbolz changed the topic of #pypy to: #pypy PyPy, the flexible snake | IRC logs: and | so many corner cases, so little time
otisolsen70 has joined #pypy
otisolsen70 has quit [Remote host closed the connection]
otisolsen70 has joined #pypy
<mattip> The release is out. :)
<ronan> nice!
<cfbolz> mattip: awesome, will post a tweet
<mattip> thanks
<cfbolz> mattip: probably not a good idea to fix it, but the blog url is wrong :-(
<mattip> I copied the last release note file from 2022/02 to 2022/03 without changing the name :(
<mattip> will try to do better in the future
<mattip> github pages does not have a .htaccess that would allow a server-side redirect, and creating a duplicate file would mess up the indexing/TOC
<arigato> thanks for the release!
<mattip> :)
<cfbolz> mattip: yes, agreed, next time
<arigato> also, any reason the release number is 7.3.9? it has been 7.3.x since 2020
<cfbolz> arigato: because we are binary compatible and it means you can reuse all your compiled extensions
<arigato> ah, OK, there *is* a reason, however strange
<arigato> "pypy: the project whose version number always starts with 7.3"
<cfbolz> :-)
<cfbolz> at some point we will stop being compatible
* cfbolz blames setuptools
otisolsen70 has quit [Quit: Leaving]
<cfbolz> ronan: thanks for doing the ast rewrite backport
<cfbolz> I should probably have done it on the default in the first place
Dejan has quit [Remote host closed the connection]
otisolsen70 has joined #pypy
<cfbolz> mattip: I am merging speedup-dict-copy now
<ronan> cfbolz: well, I should have implemented it at interplevel in the first place...
<cfbolz> :-)
<ronan> I'll be glad to remove a dependency on long-dead pytest internals
<cfbolz> yes, agreed
<cfbolz> assert reinterpretation is a mess
<arigato> mattip: one of the lines on checksums.html is indeed shorter by one character
<arigato> fc62f2c453fb851d10a1879c6e75c31ffebbf2a44d181bb06fcac4750d023fc pypy3.7-v7.3.9-aarch64.tar.bz2
<mattip> ahh, thanks. missed that
Dejan has joined #pypy
* mattip adding a test
<ronan> cfbolz: actually, reinterpretation wasn't used for apptest files, so the backport doesn't change anything. But I guess we could now use rewriting for AppTest classes as well
<cfbolz> ronan: yes, I didn't do that part yet
<mgorny> cfbolz: thanks for merging my secfixes. however, fwics two fixes for 2.7 are still missing
<mgorny> could you double-check?
<mgorny> (not claiming they're somewhat urgent given limited use of py20
<mgorny> py2)*
<cfbolz> ouch :-(
<cfbolz> mgorny: yes, I had assumed that we had caught up to all the 2021 fixes already
<cfbolz> annoying, given that we just released, my mistake, I apologize
<mgorny> no problem
<mgorny> in other news, i've hit:
<mgorny> File "/tmp/portage/dev-python/pypy-7.3.9/work/pypy2.7-v7.3.9-src/lib-python/2.7/test/", line 78, in test_negative_crc_iv_
<mgorny> AssertionError: 153214308 != -153214364
<mgorny> self.assertEqual(zlib.crc32('ham', -1), zlib.crc32('ham', 0xffffffffL))
<mgorny> input
<mgorny> will verify in a minute if it's zlib upgrade problem or not
lritter has joined #pypy
Atque has joined #pypy
<mgorny> yep, looks like it
<mgorny> let's check cpython
<cfbolz> mgorny: ouch, what
<cfbolz> Ah, some signedness thing
<mgorny> probably
<mgorny> happens with zlib-1.2.12
<mgorny> looks like some major changes happened in the crc code
<mgorny> hmm, apparently the test for negative IV was added to py2 in 2008, fwics early 2.6 alpha
<mgorny> i don't think it was even part of py3k
<mgorny> ever*
<mgorny> ok, i think i have enough data to file a bug ;-)
<cfbolz> mgorny: thanks for digging
<mgorny> lemme just rebuild a bunch of cpythons, in case it affects them after rebuilding against new zlib
<mattip> it seems we are using zlib 1.2.7 in the manylinux2014 images
<mgorny> mattip: 1.2.12 has a vulnerability fix, so you probably want to upgrade ASAP
<mgorny> otoh it seems that the vulnerability has been reported in 2018.. and they've only done the release now
<mattip> well there is from 2005 ...
<mattip> ahh, no, that was way before 1.2.7
<mattip> we don't ship it, we only use it. So the problem is not really ours, except if we need to change the way we consume it
<mgorny> ah, ok
<mgorny> cfbolz, mattip: here's the bug
<mgorny> so fwics cpython is not affected
Atque has quit [Quit: ...]
<mgorny> there's also some regression in test_multiprocessing, i'll report it in a moment
Atque has joined #pypy
Atque has quit [Quit: ...]
<cfbolz> mattip: I merged speedup-dict-copy
Hodgestar has quit [Ping timeout: 260 seconds]
Hodgestar has joined #pypy
greedom has joined #pypy
greedom has quit [Remote host closed the connection]
greedom has joined #pypy
slav0nic has joined #pypy
slav0nic has quit [Remote host closed the connection]
slav0nic has joined #pypy
lritter has quit [Quit: Leaving]
<mattip> cool
greedom has quit [Remote host closed the connection]
epony has quit [Quit: QUIT]
otisolsen70 has quit [Quit: Leaving]
slav0nic has quit [Ping timeout: 260 seconds]
Atque has joined #pypy