IRC channel logs

<dsmith_>sneek, botsnack

<sneek>:)

<sneek>Welcome back dsmith_!

<dsmith_>sneek, botsnack

<sneek>:)

<wingo>moo

<cow_2001>wingo: 無

<lloda>wingo: do you remember what's the '10' in jit.c:add_pending_reloc() ?

<wingo>lloda: i think it is an initial size. the buffer doubles as needed but doubling zero is zero

<lloda>oh ok

<lloda>i found a couple 10s that should be SCM_GSUBR_MAX, but i couldn't tell about that

<lloda>thx

<wingo>might as well have done size = size * 2 + 10, would be more normal :P

<lloda>right :-)

<chrislck>anyone found that adding & using (ice-9 match) makes compilation quite a bit slower?

<mwette>I think it depends how many cases you have. I'm guessing the CPS macro style can generate deep scoping. I have had cases where sxml-match took a long time to compile and then crashed. I fixed by breaking it up into multiple matches by making the last case "anything else" => call next sxml-match procedure.

<wingo>ice-9 match should compile in a more intelligent way in the front-end; we should do https://dl.acm.org/doi/abs/10.1145/1411304.1411311 like racket does

<rekado>does anyone here know of libusb bindings for Guile?

<RhodiumToad>FFI is your friend

<rekado>is “friend” the first or the second F in FFI?

<civodul>:-)

<civodul>maybe that's a job for nyacc?

<civodul>ACTION looks at mwette

IRC channel logs

2023-04-27.log