11:29 <whitequark> rjo: wow, you even use gpg.

11:31 <rjo> whitequark: since 1999 ;) my old key is pretty well placed in the strongly connected set circlejerk even now...

11:35 <whitequark> I see you are fond of the web of trust concept...

12:47 <whitequark> hmm, inheritance proves to be somewhat more tricky to implement than i expected...

13:00 <cr1901_modern> What's the wall for you?

13:24 <GitHub159> [artiq] jordens pushed 2 new commits to master: https://git.io/vrkhz

13:24 <GitHub159> artiq/master 4d3c734 Robert Jördens: README: link artiq from the logo

13:24 <GitHub159> artiq/master 9898daf Robert Jördens: README: DOI/citation format

13:24 <bb-m-labs> build #681 of artiq is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/681 blamelist: Robert J?rdens <rj@m-labs.hk>

13:25 <rjo> whitequark: is "just copy the methods into the namespaces" too naive?

13:25 <rjo> bb-m-labs: why?

13:25 <rjo> oh

13:26 <whitequark> rjo: there are no namespaces

13:26 <whitequark> every function is piped through to the backend

13:26 <whitequark> and the fact that the function happens to have a self argument is completely irrelevant

13:28 <GitHub139> [artiq] jordens pushed 1 new commit to master: https://git.io/vrkjJ

13:28 <GitHub139> artiq/master f8c38c1 Robert Jördens: setup.py: utf-8

13:28 <bb-m-labs> build #682 of artiq is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/682 blamelist: Robert J?rdens <rj@m-labs.hk>

13:29 <rjo> whitequark: ok. then "just copy the functions"?

13:29 <whitequark> copy them when/

13:29 <whitequark> *?

13:30 <whitequark> I mean, I know the answer. when a method object is created. but that's nontrivial to plumb through because none of the compiler was designed to make it easy...

13:32 <GitHub40> [artiq] jordens pushed 1 new commit to master: https://git.io/vrkjy

13:32 <GitHub40> artiq/master 47f3e70 Robert Jordens: setup.py: really utf-8

13:33 <rjo> whitequark: ack.

13:37 <whitequark> in other words, the compiler does not specially treat methods in any way right now, which leads to straightforward code and matches python semantics

13:37 <whitequark> this inheritance thing requires me to add a bunch of special cases on top of that.

13:38 <whitequark> there are all sorts of fun questions to answer, like "what happens if someone calls a method directly"

13:38 <whitequark> I mean cls.method(instance, ...)

13:41 <rjo> yes. that is something i'd be happy to disallow (if it help).

13:41 <whitequark> it... doesn't help with implementation, in fact it just adds a bunch more special cases

13:42 <whitequark> it might help by reducing confusion

13:47 <cr1901_modern> How does "normal Python" handle inheritance then if methods are "just another type"?

13:47 <bb-m-labs> build #417 of artiq-kc705-nist_clock is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-kc705-nist_clock/builds/417

13:53 <bb-m-labs> build #173 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/173

13:55 <bb-m-labs> build #683 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/683

19:33 <cr1901_modern> sb0: ping whenever you have free time/are awake. I'm having trouble with a MiSoC design and am unsure what's going on other than Xilinx insists on optimizing away all sequential logic (despite me having a clock connected at top level)