sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
<sb0> _florent_, why is there a socket in here?!
<sb0> is that the residue of another horrid hack in asyncio?
<sb0> whitequark, is the win7-experimental vm dead_
<sb0> ?
<whitequark> sb0: no. it's just not booted up. i told you already how to do it...
<whitequark> well, i booted it/
<sb0> the answer to my question is yes
<sb0> (re. asyncio)
<sb0> loop.run_until_complete(asyncio.sleep(1)) creates a *tcp socket*
<sb0> when i get bored, i'll rewrite asyncio correctly ...
<whitequark> _florent_: what's wrong with C99 exactly?
<whitequark> a compiler that doesn't support a 17-year-old standard is not worth bothering with
<sb0> this socket is apparently used to support communication with threads
<whitequark> yes. that's fairly typical.
<sb0> with TCP/IP?
<sb0> some xilinx tools even use sunrpc for this ...
<whitequark> not with TCP/IP normally, no. pipes support IOCP.
<sb0> i wonder if they did that because the asyncio code is littered with socket-specific code all over the place...
<sb0> hm, doesn't seem so
<sb0> oh, wait, anonymous pipes are not overlapped on windows
<sb0> maybe "reopening" them can put them in overlapped mode, but I'm not sure
<whitequark> nope.
<whitequark> Microsoft itself suggests emulating anonymous overlapped pipes with named pipes: https://github.com/ijprest/presto-build/blob/master/pipeex.c?ts=2
<sb0> yes, they do that in another place of asyncio
<sb0> windows api is crap
<whitequark> better than POSIX
<sb0> in that instance no, you can use fcntl() to set nonblocking mode at anytime on a fd
<whitequark> nonblocking mode != overlapped. POSIX only gives you select() to handle nonblocking mode fds.
<whitequark> well, of course there's POSIX AIO, but it doesn't actually work anywhere
<whitequark> something like epoll+signalfd+timerfd works. or kqueue. kqueue is pretty good actually
mazzoo has quit [Ping timeout: 260 seconds]
mazzoo has joined #m-labs
mazzoo has quit [Ping timeout: 264 seconds]
mazzoo has joined #m-labs
<GitHub8> [misoc] enjoy-digital pushed 1 new commit to master: https://git.io/v2bND
<GitHub8> misoc/master 78ce2d6 Florent Kermarrec: software: use gnu99 standard for GCC and revert liballoc changes
<whitequark> _florent_: thanks
<_florent_> whitequark: nothing wrong with C99, I'm using an old pre-compiled GCC on windows and it was not the default standard, I reverted changes in liballoc and forced gnu99 standard with GCC
FabM has joined #m-labs
rohitksingh has joined #m-labs
<GitHub25> [artiq] sbourdeauducq pushed 4 new commits to master: https://git.io/v2bjO
<GitHub25> artiq/master 6837160 Sebastien Bourdeauducq: test: I2C/PCA9548 unittest
<GitHub25> artiq/master 125ab3e Sebastien Bourdeauducq: runtime/i2c_read: fix MSB
<GitHub25> artiq/master 36387af Sebastien Bourdeauducq: examples/device_db: add PCA9548
<GitHub89> [misoc] enjoy-digital pushed 5 new commits to master: https://git.io/v2Neg
<GitHub89> misoc/master aed8829 Florent Kermarrec: interconnect/stream: remove sop support
<GitHub89> misoc/master 4bf472e Florent Kermarrec: cores/liteeth_mini: remove use of sop...
<GitHub89> misoc/master 5f124ec Florent Kermarrec: cores/liteeth_mini/phy: remove with_hw_init_reset parameter...
rohitksingh has quit [Ping timeout: 260 seconds]
rohitksingh has joined #m-labs
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined #m-labs
rohitksingh has quit [Quit: Leaving.]
FabM has quit [Remote host closed the connection]
<GitHub71> [misoc] enjoy-digital pushed 6 new commits to master: https://git.io/v2N2q
<GitHub71> misoc/master e4868b9 Florent Kermarrec: interconnect/stream: remove packetized parameter from EndpointDescription...
<GitHub71> misoc/master 1dff22b Florent Kermarrec: cores/liteeth_mini/phy: remove pads_register parameter (does not save enough logic to be kept)
<GitHub71> misoc/master bce6c5c Florent Kermarrec: cores/liteeth_mini/common: remove Port (was only used once)
mazzoo has quit [Remote host closed the connection]