11:44 <hl> whitequark: The USB4 specification has been published. https://usb.org/document-library/usb4tm-specification

11:46 <tnt> hl: https://twitter.com/whitequark/status/1168936985316265984

11:46 <hl> ah okay

11:46 <hl> I've been away for a few days so

11:48 <hl> ...jesus christ wtf

12:12 <gruetzkopf> it's interesting..

12:13 <gruetzkopf> cursed-interesting

12:13 <hl> yeahhhh

12:13 <hl> so if I'm reading this right, USB4 hubs are required to be PCIe-aware and have a PCIe switch despite PCIe being tunneled. WTF?

12:14 <hl> not only that, the diagram almost seems to suggest they think the PCIe switch in a hub is going to be a separate chip... with both its upstream and downstream ports connected to the hub chip. WTF?????

15:15 <ExeciN> Hi people. I was looking to get a logic analyser and I've stumbled upon Glasgow. Has anyone experimented with any kind of software + glasgow specifically for logic analysis?

15:36 <tnt> ExeciN: there is no such support yet

15:38 <ZirconiumX> "yet" being the key word here

16:49 <gruetzkopf> hey, the current HEAD commit doesn't work for me, running any applet shows output close to this: https://paste.ccc.ac/?db9f7578454078bc#ndXx8KUCWz8tYLgD/6vgwhRgGS3IuJcH+UfJtcOuqME=

17:13 <emily> weird

17:13 <emily> what's in build_top.sh? or does it clean that up too well to take a look at it

17:13 <emily> I wonder if the _toolchain stuff might be to blame here

17:13 <emily> also I assume you're using latest nmigen right?

17:15 <gruetzkopf> i was using exactly what glasgows setup process pulled in, nmigen HEAD doesn't change output except for commit id

17:17 <emily> yeah. well it's hard to know what's up beyond a syntax error in the build_top.sh script

17:17 <emily> (which I guess is >700 lines long somehow??)

17:18 <emily> dunno where it puts those, maybe in /tmp somewhere?

17:18 <emily> wq likely has a better idea ofc, but I touched this part of nmigen recently so maybe I can help ^^;

17:31 <gruetzkopf> no clue where that file goes

17:33 <gruetzkopf> oh, so it creates a temporary directory, but i'm too slow to catch that

17:34 <emily> you could just s/sh/cat/ in nmigen/build/run.py:98

17:34 <emily> I feel like it should leave around the files for inspection in case of error though...

17:37 <jn__> or sh -x maybe?

17:40 <emily> hm, would be nice if it used sh -x always, buffered the output, and printed it in case of error

17:44 <gruetzkopf> added a copy to that location, https://paste.ccc.ac/?6b69ce203aa2c776#ATYPhvBDV4PUKmMLdj52uJpVVOaj2BK8+9rKMbw1fSI= output looks reasonable

18:14 <gruetzkopf> current yosys - no change

18:17 <emily> oh, I guess the syntax error is coming from yosys or something rather than the shell...

18:20 <emily> should have been obvious in retrospect, sorry ^^;

19:31 <gruetzkopf> not getting any output from any subtool

19:32 <gruetzkopf> yosys, nextpnr-ice40 and icepack are all functional

19:57 <emily> so uh... does the script work if you manually generate the build directory and then run build_top.sh from there, or?

22:05 <whitequark> gruetzkopf: this looks like outdated yosys

22:05 <whitequark> has nothing to do with build_top.sh script, it means yosys can't read generated RTLIL

22:08 <emily> hm, I thought nmigen checked yosys version though?

22:13 <whitequark> it does

22:13 <whitequark> not sure why it broke here

22:19 <emily> curious to see gruetzkopf's `yosys -V`