09:58 <pepijndevos> pucteznxrwd

09:58 <pepijndevos> Maybe not EW??

10:00 <trabucayre> 10:31 <@pepijndevos> pucteznxrwd <- cat walking on keyboard ?

10:01 <pepijndevos> just my hands...

10:04 <trabucayre> I have thinking to a password too :-)

10:04 <pepijndevos> related: https://twitter.com/eggplantbear/status/1110248100172423168

10:14 <pepijndevos> Blinky is also using EW an NS wires and works so maybe it's not that...

10:16 <trabucayre> It's really interesting, I think I understand words but the result is not really clear :-(

10:24 <pepijndevos> So what I'm doing is looking at nextpnr JSON ROUTING attribute. And then I'm tracing the route from the reset IBUF to the LED OBUF. And for each step I try to see if it matches with the unpacked verilog.

10:25 <pepijndevos> Yesterday it appeared that there was R2C6_EW20 in the Nextpnr JSON from the reset inverter LUT to the SR port of the flipflop that was missing in the unpacked code

10:27 <pepijndevos> That would mean that either there is a bug in the unpacker, a human error in my debugging, a bug in the packer that didn't set this pip correctly, or a bug in nextpnr that incorrectly wired something. So I'm trying to rule out these options

10:38 <pepijndevos> Ok maaaaybe there is a more obvious problem here...

10:39 <pepijndevos> unpacker only sees DFF(E) while yosys output has DFFP and other variants

10:39 <pepijndevos> omg

10:39 <pepijndevos> that would explain why blinky does work, but async reset on shift does not.

10:40 <pepijndevos> how am I even preserving flop type in nextpnr... I don't remember... what If I don't....

10:42 <pepijndevos> OMG the packer just doesn't consider it

10:46 <pepijndevos> YESSSSS

10:52 <pepijndevos> trabucayre, should I add your shift example to the folder? It's apparently a good test for different flip-flop types

10:55 <pepijndevos> well fuck... attosoc still doesn't work

10:57 <pepijndevos> trabucayre, do you have any more of those nice broken progams? XD

11:09 <trabucayre> Yes of course You can :)

11:09 <trabucayre> will retry my runber clock :)

11:13 <pepijndevos> I guess now is a good time to update the Tang Nano LCD example to see if that one works.

11:13 <pepijndevos> That's also a moderately sized one that is... potentially debuggable

11:42 <trabucayre> It's amazing when reading the commit to see the simplicity... But this hide the complexity to find issue

12:45 <pepijndevos> :'( yes

14:13 <trabucayre> pepijndevos: maybe will try PMOD demo

14:31 <pepijndevos> ah you have the oled pmod?

14:31 <pepijndevos> that would be interesting

14:38 <trabucayre> maybe the same (I'm not sure)

14:52 <pepijndevos> Tempted to buy some pmods for my collection of stuff I never use

14:53 <pepijndevos> Or maybe I should just finally take the time to do something with the audio pmod which I bought with a project in mind

15:19 <trabucayre> I'have some stupid thing with pmod I2S :)