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
fryguybob has quit [Ping timeout: 276 seconds]
fryguybob has joined #pypy
commandoline has quit [Remote host closed the connection]
commandoline has joined #pypy
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
wleslie has joined #pypy
adamholmberg has joined #pypy
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
jvesely has joined #pypy
adamholmberg has quit [Remote host closed the connection]
lritter has quit [Ping timeout: 265 seconds]
adamholmberg has joined #pypy
adamholmberg has quit [Remote host closed the connection]
jcea has quit [Quit: jcea]
<bbot2> Failure: http://buildbot.pypy.org/builders/rpython-win-x86-32/builds/218 [mattip: force build, release-pypy3.6-v7.x]
junna has joined #pypy
junna has quit [Ping timeout: 240 seconds]
dmalcolm_ has joined #pypy
dmalcolm has quit [Ping timeout: 265 seconds]
<mattip> The x64 docker buildslave uses gcc 8.3. rpython asmgcroot tests are failing
<mattip> I think the problem is in assembler like this
<mattip> with .LHOT and .LCOLD labels
<mattip> I compared the *.s file to gcc 7.4, that was the biggest change
<mattip> and that label appears in the iterline failure in rpython/translator/c/gcc/trackgcroot.py:find_functions
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/1199 [mattip: force build, release-pypy2.7-v7.x]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/1200 [mattip: force build, release-pypy3.6-v7.x]
<mattip> definitely gcc-8 related, when I run the test on Ubuntu 18.04 with "CC=gcc-8" it fails
<mattip> another option is to remove trackgcroot
<wleslie> could build with the shadowstack until we can update asmgcroot?
<wleslie> or replace it with something maintainable (like something DWARF based)
<mattip> this is only in tests. We don't use this in pypy
<wleslie> oh
<mattip> well, we don't use it *anymore*
<wleslie> what's the current rootfinder?
<mattip> shadowstack, asmgcc is destined to be depricated https://doc.pypy.org/en/latest/config/translation.gcrootfinder.html
<mattip> maybe now is the time :)
<wleslie> I wholeheartedly approve. Which is a first for me on removal of code.
_whitelogger has joined #pypy
junna has joined #pypy
<arigato> mattip: yes, I agree to kill all related code if it starts failing tests
<kenaan> mattip pypy.org[extradoc] 957dc5e9e470 /source/download.txt: add pypy3.6 7.3.0rc1 hashes
<mattip> 7.3.0rc1 release candidates are up
<arigato> yay!
<mattip> antocuni was having problems with curses, I hope they have gone away now with ncurses6
<arigato> mattip: I can kill asmgcroot-related codes myself today
<mattip> +1
<mattip> there is scattered mentions in documentation and configuration too
tsaka__ has quit [Quit: Konversation terminated!]
tsaka__ has joined #pypy
<cfbolz> mattip: I had discussed it with armin, we should kill it
<cfbolz> ah, sorry, didn't read far enough :-)
<arigato> :-)
<tos9> (FWIW, GH Actions has PyPy{2,3} latest as of yesterday hooray)
<cfbolz> nice
<tos9> cfbolz: also hello :)
<tos9> how goes
<cfbolz> pretty good, new house and new baby :-)
tsaka__ has quit [Read error: Connection reset by peer]
<cfbolz> you?
<tos9> congrats! figured it was about that time
<cfbolz> thanks
<tos9> all good here, in the EU for the next few weeks and not working, so those are both good things
<cfbolz> nice! whereabouts?
<tos9> I'm in Portugal at the moment but we're probably going to drive to spain in the next few days
<cfbolz> awesome
<tos9> yeah pretty nice here, 16 and reasonably sunny is not bad for this time of year...
<cfbolz> beach weather by german standards :-P
<tos9> ha
junna has quit [Ping timeout: 240 seconds]
* arigato is in 1 meter of snow
<kenaan> arigo kill-asmgcc b23b24b53574 /: in-progress: remove asmgcc completely (modern versions of gcc show test failures)
<mattip> rc1 downloads are at https://bitbucket.org/pypy/pypy/downloads
tsaka__ has joined #pypy
jvesely has quit [Quit: jvesely]
lritter has joined #pypy
senyai has quit [Quit: Leaving]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6652 [arigo: force build, kill-asmgcc]
<bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7871 [arigo: force build, kill-asmgcc]
marky1991 has joined #pypy
marky1991 has quit [Remote host closed the connection]
marky1991 has joined #pypy
tsaka__ has quit [Ping timeout: 240 seconds]
<bbot2> Started: http://buildbot.pypy.org/builders/rpython-linux-x86-64/builds/250 [mattip: force build, kill-asmgcc]
<mattip> arigato: the rpython tests are now run separately ^^^
marky1991 has quit [Ping timeout: 268 seconds]
junna has joined #pypy
<bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7871 [arigo: force build, kill-asmgcc]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/6652 [arigo: force build, kill-asmgcc]
CountryNerd has joined #pypy
tsaka__ has joined #pypy
tsaka_ has joined #pypy
tsaka__ has quit [Ping timeout: 246 seconds]
<arigato> mattip: ah right, thanks
<kenaan> arigo kill-asmgcc 04030975247c /pypy/doc/whatsnew-head.rst: Document branch
CountryNerd has quit [Quit: CountryNerd]
junna has quit [Ping timeout: 268 seconds]
jcea has joined #pypy
<arigato> a python gotcha: if d is a dict, the line "d = d[key] = {}" doesn't do what you think
<Alex_Gaynor> arigato: pfff, what do I think it does :-)
<arigato> it's meant to do "d1 = {}; d[key] = d1; d = d1"
<arigato> except it's doing something different
<arigato> better just write it explicitly
<Alex_Gaynor> yes, indeed.
<arigato> what it *does* is "d = {}; d[key] = d"
<arigato> which is a bit unexpected
<arigato> I guess I could write it like this: d[key] = d = {}
<arigato> that's very high on the obscure side though
<Alex_Gaynor> it's all obscure :-) in C it would be undefined behavior to both write and read to one variable without a sequence point :P
<mattip> So the assignments in `a = b = c = d = None` are done left to right, not right to left?
<arigato> yes
<mattip> dis.dis agrees, first a, then b, then c, then d using a DUP_TOP bytecode to duplicate the None value
<arigato> it's at least consistent with: e = {}; d, d["key"] = e, e
<arigato> which also evaluates the d["key"] with the value of d already modified to be equal to e
tsaka__ has joined #pypy
tsaka_ has quit [Read error: No route to host]
<bbot2> Failure: http://buildbot.pypy.org/builders/rpython-linux-x86-64/builds/250 [mattip: force build, kill-asmgcc]
adamholmberg has joined #pypy
jvesely has joined #pypy
junna has joined #pypy
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
marky1991 has joined #pypy
adamholmberg has quit [Remote host closed the connection]
marky1991 has quit [Ping timeout: 268 seconds]
adamholmberg has joined #pypy
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
idnar has quit []
idnar has joined #pypy
dddddd has joined #pypy
avakdh has quit []
avakdh has joined #pypy
stillinbeta has quit []
stillinbeta has joined #pypy
adamholmberg has quit [Remote host closed the connection]
marky1991 has joined #pypy
junna has quit [Ping timeout: 268 seconds]
i9zO5AP has joined #pypy
Ai9zO5AP has quit [Ping timeout: 250 seconds]
dddddd has quit [Ping timeout: 240 seconds]
<ronan> arigato: did you have a look at https://bitbucket.org/pypy/pypy/pull-requests/686 ?
dddddd has joined #pypy
avakdh has quit []
stillinbeta has quit []
avakdh has joined #pypy
stillinbeta has joined #pypy
adamholmberg has joined #pypy
adamholmberg has quit [Remote host closed the connection]
lazka has joined #pypy
<lazka> mattip, in case it hasn't been reported, SSL is broken for the rc release. On Ubuntu at least
<mattip> how is it broken?
<mattip> which ubuntu?
<lazka> Ubuntu 19.10
<mattip> thanks for the report
<lazka> the squeaky builds work
<mattip> hmm, can reproduce on 18.04 as well :(
<mattip> for the logs: ./pypy3 -c "import urllib.request; urllib.request.urlopen('https://www.google.com')"
<mattip> fails with <urlopen error [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed>
<mattip> probably we need to apply the ssl patches from portable-pypy
<mattip> lazka: thanks, indeed the portable-pypy patches fix it
<lazka> all my other tests work as before (only PyOS_FSPath() still missing)
<lazka> ok, cool :) bye
lazka has quit [Remote host closed the connection]
<mattip> thanks. The file squeaky patches the CPython ssl.py. I wonder if we should be fixing the cffi _ssl.py instead
jvesely has quit [Quit: jvesely]
lazka has joined #pypy
jcea has quit [Remote host closed the connection]
<lazka> mattip, I think the SSL_CERT_FILE/SSL_CERT_DIR are also build time constants in openssl, so it would make sense that a portable build would ignore them
jcea has joined #pypy
<lazka> while a system build wouldn't
lazka has quit [Remote host closed the connection]
<arigato> ronan: if issue 3096 is still fixed with your pull request, I have nothing against it
marky1991 has quit [Read error: Connection reset by peer]
<arigato> I guess I'm not understanding why it fails CPython's own tests given that the code you revert tries to follow more closely what CPython does
<arigato> but I'm fine with the pull request anyway
<arigato> it's typically the kind of thing where CPython and PyPy would always differ very subtly, just because what CPython really does is unexplainable in more abstract terms
Kipras_ has joined #pypy
ljwobker has joined #pypy
ljwobker has quit [Remote host closed the connection]
<ronan> arigato: well, in my understanding, the code before your changes was actually closer to CPython, since it doesn't have an error stack
Kipras_ has quit [Ping timeout: 240 seconds]
jcea has quit [Remote host closed the connection]
jcea has joined #pypy
lritter has quit [Quit: Leaving]
tsaka__ has quit [Read error: Connection reset by peer]