<Wanda[cis]>
162 files changed, 12011 insertions(+), 8672 deletions(-), 343 build errors left to go
<Wanda[cis]>
this refactor is going about as well as I expected
DemiMarieObenour has joined #prjcombine
<DemiMarieObenour>
What about just writing a code generator?
<Wanda[cis]>
I was considering it (or, well, a proc-macro which is the same thing), but in the end the plain macro did the trick
<Wanda[cis]>
it's still a little wonky, but there's a good chance that applying mei's suggestions will fix it, which I'll try out when half my tree isn't completely broken
<Wanda[cis]>
generating the list of bel slots is only a tiny part of this refactor, completely changing the model of how bels are identified is what actually generates the giant diffstat by necessity
<Wanda[cis]>
also I got so annoyed at trying to convert ise-hammer last night that I decided to just rewrite it, basing the new version on the much cleaner xact-hammer
<Wanda[cis]>
I have made it worse (added more traits to share some code between xact-hammer and the new ise-hammer)
cr1901 has quit [Read error: Connection reset by peer]
cr1901 has joined #prjcombine
<Wanda[cis]>
alright the new ise-hammer core is working (has reversed the S3 DSP tile)
<Wanda[cis]>
now I just need to port the entire rest of the thing, brb.