MasterDuke .tell jnthn "has been to use IO::Scoket::Async" 00:10
yoleaux MasterDuke: I'll pass your message to jnthn.
Geth MoarVM: 4304fb3dd5 | (Samantha McVey)++ | 2 files
[unicode-grant] Add string eqat/index ignorecase+ignoremark funct's

To add this function with a minimum extra code, we alter the string_equal_at_ignore_case_INTERNAL_loop to take an extra argument:
  'ignoremark'. If set to 1 it will use the new ord_getbasechar function
to search ignoring unicode marks in addition to ignoring the case.
I factored code from MVM_string_ord_basechar_at into the new ord_getbasechar function so we can reuse code between that and the INTERNAL_loop function.
I factored out similarly for MVM_string_equal_at_ignore_case and ignore_mark.		 00:35
MoarVM: ad6ab26fa6 | (Samantha McVey)++ | 6 files
[unicode-grant] Add new eqaticim_s and indexicim_s ops ignorecase+ignoremark

Adds ops for the two external functions that were added in the previous commit.
brrt morning 08:03
ooh new jnthnpost
nwc10 jnthn: IO::Scoket::Async 08:06
jnthn Grrr, 2FA is great until for some reason the code isn't sent :/ 09:00
yoleaux 00:10Z <MasterDuke> jnthn: "has been to use IO::Scoket::Async"
nwc10 I hope this wasn't a coffee machine with 2FA 09:01
jnthn On, darn wordpress 09:02
Ah well, the typos can wait then 09:03
samcv jnthn++ on the blog post 09:12
Geth MoarVM: f1c18cc3fa | (Samantha McVey)++ | src/strings/ops.c
Make string_index_ignore_case and ord_getbasechar static

  @zhuomingliang++		 09:13
jnthn Oh goodness JVM, can I really not give your darn process spawing stuff a descriptor to inherit :/ 09:15
*for the process to inherit 09:16
samcv night all o/ 09:29
jnthn o/ samcv
brrt anyway, i've been thinking and thinking and thinking 11:03
the tl;dr
the least complex way to go forward is to take out point spilling
it does not interact well with the retroactive action of the full spilling 11:04
which is interesting
because linear scan moves forward, so you can usually assume that your local decisions are not invalidateed
but spillling also works backwards
the other interesting bit is 11:07
i currently don't insert loads for ARGLIST refs 11:08
because arglist can have more refs than can possibly fit in regsiters
but obviously, that's not going to be an issue for retroactively-spilled arglist refs
since, well, they must have fit for the value to have been live at that point 11:09
samcv good ???? 18:44
Geth MoarVM: 54d999b330 | (Samantha McVey)++ | src/strings/ops.c
Make string_equal_at_ignore_case static

  @zhuomingliang++		 21:44

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!