00:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/875 [default]

00:12 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/3827 [default]

00:17 <bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1802 [default]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/3828 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5530 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/6808 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3963 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/5912 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4679 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armel/builds/2322 [py3.5]

01:00 <bbot2> Started: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armhf-raspbian/builds/2083 [py3.5]

02:00 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/4679 [py3.5]

02:11 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5530 [py3.5]

02:15 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armel/builds/2322 [py3.5]

02:21 <bbot2> Success: http://buildbot.pypy.org/builders/build-pypy-c-jit-linux-armhf-raspbian/builds/2083 [py3.5]

02:21 <bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-armhf-raspbian/builds/1775 [py3.5]

02:23 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/6808 [py3.5]

04:15 <bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-32/builds/5912 [py3.5]

04:40 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/3828 [py3.5]

04:50 <mattip> it seems randomly trying to fix unicode-utf8-py3 without a deeper understanding is rather useless

04:52 <mattip> space.unicode_w still exists on baseobjspace, should it go away?

04:52 <mattip> sometimes we can accept utf8 with surrogates, when and why?

04:54 <mattip> I thought the whole point of this branch was that we do not use unicode internally, only utf8, true or false?

04:54 <mattip> the places that do decode_utf8 on py3.5, do they now do str_decode_utf8 or something else?

04:55 <mattip> the main problem right now is that the branch is totally broken, I don't even know where to start to fix things

04:56 <mattip> </rant>

05:32 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/3963 [py3.5]

06:14 <fijal> mattip_: the 2 version completely removed Unicode internally

06:15 <fijal> I do not know why the 3 one didn't, maybe for reasons like "let's run it"

08:12 <twn> Hi I have a question about using rffi. Say I have a C function prototype like this: void mock_print(long* data, unsigned long len);

08:13 <twn> How do I hook that up using rffi.llexternal?

08:13 <twn> I feel like it should be something roughly like this: mock_print = rffi.llexternal("mock_print", [lltype.Array(lltype.Signed), lltype.Unsigned], lltype.Void, compilation_info=info)

08:14 <twn> That isn't working (and may be deranged), but I'm not sure how to call this

08:15 <twn> As a secondary question, how would this then be called? I.e. I'd like to do something like this: mock_print([1, 2, 3], 3) but I'm not sure if that conversions are done automatically

08:15 <twn> Thanks a lot in advance for any help!

08:16 <cfbolz> mattip_: yes, imo unicode_w should go away, only bytes_w, text_w and utf8_w should be left

08:22 <fijal> twn: you definitely want lltype.Ptr

08:23 <fijal> but also lltype.Array is an RPython array, not c level something*

08:23 <fijal> so you want rffi.CArrayPtr(lltype.Signed)

08:23 <fijal> the conversions are not automatically, partly because of the lifetime considerations

08:24 <fijal> so you would need to create your own lltype.malloc(rffi.CArray(lltype.Signed), 3, flavor='raw')

08:24 <fijal> and then lltype.free it

08:24 <fijal> there might be some helpers though

08:34 <twn> great thanks just tested it out and it works perfectly! thanks for clearing up the confusion!

08:35 <fijal> I'm sorry it's not that well documented

08:46 <twn> no worries! i'm having a little fun twisting my brain around into understanding how things hang together with pypy and rpython so a little struggle isn't totally unwelcome :)

08:47 <twn> besides maybe once i sort things out and have the confusion fresh in my mind i can make some improvements to documentation (no promises!)

10:02 <twn> Hi I have another question. Am I able to use descriptors in rpython? I have a custom __getitem__ defined, but it throws an error like "[translation:ERROR] AttributeError: 'FrozenDesc' object has no attribute 'getuniqueclassdef' "

10:02 <twn> Here is a detailed gist: https://gist.github.com/ApproximateIdentity/c34af60a6657f391aee10f84d28a2c6b

10:05 <cfbolz> twn: most special methods don't work. getitem does but slice is not supported

10:06 <twn> ah okay thanks

13:41 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-s390x/builds/875 [default]

14:31 <twn> Hello I was wondering if anyone wants to let me pick their brain a bit. I could also take this to the mailing list, but I figure I would ask here first. I started working on this issue last week: https://bitbucket.org/pypy/pypy/issues/1901 The point as I understand it is to change the code in rpython/rlib/rrandom.py to use compiled binary code to see if it speeds things up. I decided to try it out mainly to see if I could learn a bit more about

14:31 <twn> pypy and as a by-product maybe doing something useful. So what I've done so far is to first take the code in numpy and adapt it to using cffi as an installable module to pass the tests. Then I changed that to use rffi and compiled a small binary in rpython which is also passing the tests. Then I moved that code into the interpreter itself and got it to build. The problem at this point is that I realized that I am using llexternal to call code

14:31 <twn> that I feel should just be directly compiled into the interpreter instead of my current method which runs into issues. Anyway is this to sort of thing to bug people on irc about or the mailing list or the issue tracker or ... ? In any case, thanks a lot for any help and advice!

14:34 <twn> btw i haven

14:35 <twn> btw i haven't updated that issue i linked with the latest stuff i've done. i thought i'd ask here before continuing to write stuff there in case other areas are more trafficed

14:38 <cfbolz> twn: asking here is the right approach

14:38 <cfbolz> I'll try to take a look at your code soon, maybe will manage tonight

14:40 <twn> oh great thanks a lot. in that case i think i will update that issue with a clearer explanation of what i've done so far and the final problems i've run into. at least so that my issue comments are up to date with my code

14:40 <cfbolz> Cool

14:40 <cfbolz> twn: llexternal is the right approach imo

14:43 <twn> oh okay cool. so the problem i run into is basically this. i can compile the interpreter, but when i actually try to call the methods themselves the c functions are not being found. this is why i thought it would be best philosophically the actually have the functions compiled in directly. it feels like llexternal is better suited to maybe an external c library. of course this could all just be confusion and i am presumably just using it wrong

14:43 <twn> anyway maybe don't worry now. i'll explain the issues i'm having more clearly so that it makes more sense in the issue itself

14:44 <twn> thanks for the help

15:03 <cfbolz> twn: so it's some kind of linking problem?

15:06 <twn> yeah seems like it. i assume it's just that i'm using it wrong though

16:00 <cfbolz> twn: that stuff can be tricky to figure out

16:11 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-armhf-raspbian/builds/1774 [default]

16:24 <twn> cfbolz: i updated the issue here: https://bitbucket.org/pypy/pypy/issues/1901 the final comment is up to date. i hopefully added in enough error messages and detail so you'll be able to see what my problem is without needing to recompile everything

16:25 <twn> and don't worry about a super fast response. there isn't any rush since i'll probably be away from the comp for the rest of the evening. thanks a lot again for taking a look at things.

18:24 <bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-armhf-raspbian/builds/1775 [py3.5]

22:00 <bbot2> Started: http://buildbot.pypy.org/builders/own-linux-s390x/builds/952 [default]