<Adrien[m]>
Anyone has an idea what could be the root cause ??
<Adrien[m]>
Tried reinstalling python and all libs and regenerating the packages etc, everything should be on sync about that (Archlinux system)... to not avail
<juri_>
Adrien[m]: your message was way too long, and did not make it to IRC.
<Adrien[m]>
ah, sorry. I'll resubmit with shorter lines.
<Adrien[m]>
So, I have a question related to prjxray's tool fasm2frames
<Adrien[m]>
The optimized fasm parser is based on some antlr4 lib that has python bindings. I would like to use it. But it fails :
<Adrien[m]>
/usr/lib/python3.12/site-packages/fasm/parser/__init__.py:30: RuntimeWarning: Unable to import fast Antlr4 parser implementation.
<Adrien[m]>
ImportError: cannot import name tags
<Adrien[m]>
Falling back to the much slower pure Python textX based parser implementation.
whitequark[cis] has joined #yosys
<whitequark[cis]>
juri_: (didn't you get a link to the full message? or do you just not want to click it)
<whitequark[cis]>
(I'm the provider of the bridge so if it's broken I'd like to understand how)
<juri_>
whitequark[cis]: I did, and i did not want to click away.
<whitequark[cis]>
gotcha
<Adrien[m]>
(btw thank you Catherine for your archive of irc logs, super useful :-) )
<whitequark[cis]>
you're welcome!
<Adrien[m]>
Is there a bridge for #f4pga too ?
<whitequark[cis]>
no, but if the ops want me to set one up i can do that
<Adrien[m]>
So I've a submission deadline this WE for a paper where I make of the tools Yosys-nextpnr-prjxray
<Adrien[m]>
This fasm parser issue makes the tool killed because out of memory, hence importance of the optimized parser
<Adrien[m]>
I'll take any remotely useful suggestion to try to make this python tool fasm2frames condescend to load the underlying antlr4 lib !
emeb_mac has joined #yosys
<Adrien[m]>
I can't find where tags is supposed to come from
<Adrien[m]>
Need to setup swap space on disk to run the supposedly signoff tool fasm2frames...
<Adrien[m]>
How ridiculous :D
<gatecat>
yeah, prjxray is a bit of a mess, to say the least
<gatecat>
I don't know quite how they managed to make fasm parsing the least efficient step of the process
<gatecat>
i don't know how much any of this is being actively worked on anymore, either, unfortunately
<gatecat>
maybe there's a better trick to generating bitstreams that I never figured out, idk