00:02 finanalyst left 09:14 [Tux] left 09:19 [Tux] joined 09:55 finanalyst joined 09:59 sena_kun joined
Geth nqp/main: 45cff7ed42 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM for complete JIT disabling on ARM architectures
10:37
rakudo/main: 633df28933 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP for complete JIT disabling on ARM architectures
10:49
11:04 finanalyst left 11:55 guifa left
lizmat tonyo ugexe looks like 360.zef.pm is down 15:10
or at least *very* slow 15:11
ok, seems to be recovering now: respons time of ~ 3 seconds, rather than 1 minute+ 15:14
Geth ¦ problem-solving: coke self-assigned Rakudo REPL’s prompt not configurable github.com/Raku/problem-solving/issues/459 16:19
rakudo/coke/ps1: 8f2efd70a2 | (Will Coleda)++ | src/core.c/REPL.rakumod
Naive RAKUDO_REPL_PROMPT

Only supports a literal string, no way to add the index number.
16:36
[Coke] I remember a recent discussion about adding some terminal codes to core - where did that end up? 16:39
lizmat patrickb might remember 16:41
patrickb The idea didn't manifest (yet). I guess the key issue to conclude is how we want to add modules to core. I.e. how are they versioned, dual-life, ... 16:44
I'm tentatively con, ab5tract is pro. 16:45
ugexe fwiw the two newest commits in github.com/rakudo/rakudo/commits/u...ore-dists/ start the process of splitting the core modules into multiple distributions (so they can be versioned separately and more easily dual-lifed) 17:08
there would still be work left trying to get e.g. `make install` to work since it would need to update the code where it installed the single core dist and updated to install multiple dists. probably need some similar smaller update for `make spectest` 17:10
lizmat I hope that won't slow down the "make install" process too much: I do that a *lot* :-)
otoh, perhaps they won't need to be installed other then on an initial build 17:11
so that they would just need to be re-precomped when the setting changes
ugexe i dont expect it would, but i also don't think make install being slightly slower to Do The Right Thing is worth fretting over
lizmat on demand
ugexe i say this running make install far less than you of course :) 17:12
lizmat also, it would be nice if one wouldn't need to recompile the "c" and "d" setting, when you only changed things in "e"
and to further this thinking: perhaps "d" and "e" setting could become "normal" modules 17:14
.oO( one can dream :-)
ab5tract patrickb: I’ve been taking too long to update the problem solving ticket. I’m afraid you and ugexe have managed to convert my opinion 17:17
tonyo that 360.zef.pm endpoint is all on AWS lambda/S3 17:18
lizmat tonyo: yes, it was just very slow about 2 hours ago
so maybe a general AWS issue? 17:19
now <1 second response time
I noticed when doing a "zef update" that hung for +1 minute
tonyo yea 17:37
Geth rakudo: ugexe++ created pull request #5756:
Split core modules into separate distributions
tonyo the postal service site was down, github, etc here for a while last night
[Coke] ooh, making the interactive-prompt method slightly more complicated is causing a crash. 18:22
(ah, found an old copy of ./install/bin/raku sitting in a repl, wonder if that was borking the make install) 18:36
Geth rakudo/coke/ps1: 3d4cfd8f81 | (Will Coleda)++ | src/core.c/REPL.rakumod
Support '\i' in the PROMPT

This allows the index to be specified.
18:38
rakudo/coke/ps1: 05bd6c2334 | (Will Coleda)++ | src/core.c/REPL.rakumod
Support \t in PROMPT

for HH::MM::SS
18:51
[Coke] I assume we'd want this to actually live in v6.e ?
lizmat that'd be difficult to do I'd think 18:52
Geth rakudo/coke/ps1: 6f5b4e2020 | (Will Coleda)++ | src/core.c/REPL.rakumod
Support \A in PROMPT

for HH::MM
18:55
rakudo/coke/ps1: 888def28b5 | (Will Coleda)++ | src/core.c/REPL.rakumod
add \a for a bell to PROMPT

reformat long lines, add comments
21:38
21:48 guifa joined 22:34 guifa left 23:16 sena_kun left