tbrowder well that is messy. i would like to see just the hot path in my code without the moarvm and such. 12:45
timo can you elaborate on what your actual goal is? 12:56
do you want to like, see performance numbers? do you want to know what different places a given function was called from?
antononcube Speaking of "performance numbes" I think there should be a dedicated Number Theory package. 13:18
timo we have the profiler where you run the program with like `raku --profile=my_profile.html -I something/something blabla.raku` and you end up with a html file you can open in your browser that has a call graph tab with a flame graph (well, icicle graph) or a table of routines that you can sort by different criteria
antononcube Some of the Number Theory functions would likely be a good source of code the check and improve "performance numbers" for. 13:19
timo there's a tool called "moarperf" that's my "new" profiler frontend which wants to have the output that you get with `--profile=my_profile.sql`, it's a little cro webapp. i'm actually not exactly sure if i've got proper releases for that at the moment 13:20
github.com/timo/moarperf/releases/...30a5ea4aae here's an AppImage that contains a rakudo inside and needs no further manual work
antononcube "Math::Sequences" has a few functions like that, but, well, it should be about sequences. I expect some of its code to be factored out.
timo oh, tbrowder what OS are you on? 13:23
tbrowder debian 13:25
timo ok cool. 13:26
tbrowder hi, timo. i just want to see the various sub calls.
i'm in the middle of some very old perl-to-raku code mixed in with ps and pdf and i got too sloppy 13:28
timo "see the various sub calls" is not quite specific enough for me to give a good answer 13:29
the Raku plugin for JetBrains IDEs, aka Comma, can "go to definition" by ctrl-clicking on subs
jdv is there a raku lsp yet? i forget. 14:58
patrickb There is one created by bscan. 15:11
antononcube Yeah -- I used it all the time in VSCode. 15:13
≈95% via Jupyter, ≈%5 with Raku text/code files. 15:14
tbrowder timo thanks, sorry for the sloppy answer. i'm making progress now. 15:21
i do want to try comma-ide again, but no time at the moment. 15:22
jdv ah cool 15:30
[Coke] releasable6: status 15:39
releasable6 [Coke], Next release in ≈16 days and ≈3 hours. There are no known blockers. Changelog for this release was not started yet
[Coke], Details: gist.github.com/6bcd86e439edcae1be...4d613e8af9
msiism [Coke]: Should the problem report about the REPL prompt go into Rakudo problems or MoarVM problems? 15:44
lizmat msiism: you mean about the index? then rakudo 15:49
msiism Yes, that. Thanks. 15:52
Geth ¦ problem-solving: coke self-assigned Rakudo REPL’s prompt not configurable github.com/Raku/problem-solving/issues/459 16:19
[Coke] Thanks. I'll poke around! 16:22
lizmat [Coke]: also, raku.land/zef:lizmat/Prompt
[Coke] Do you care if it's an env var vs. a class you plugin? 16:25
I am leaning towards env var.
jdv cli arg isnt even considered?:) 16:26
[Coke] given how big bash PS1's get, I wouldn't
jdv i like env var
msiism [Coke]: Thank you. I'll stay tuned. 16:29
I don't really know about classes you plug in, so I can't really say. But an environment variable would be fine with me. 16:30
[Coke] do we see this as a rakudo feature or a raku feature? 16:35
(putting it in with a rakudo-specific name for now)
msiism I would be able to set the environment variable from within the REPL, right? 16:36
[Coke] github.com/rakudo/rakudo/tree/coke/ps1 16:37
Yes. 16:38
msiism Good.
[Coke] you'd have do so something like [0] > %*ENV<RAKUDO_REPL_PROMPT>='abc >'
abc >
msiism I see. 16:39
lizmat note that the prompt changes with incomplete statements
jdv it could be set pre invocation too, right? 16:44
Geth ecosystem/main: 202b4b9056 | (Elizabeth Mattijsen)++ | META.list
Remove Editsrc::Uggedit

it has bitrotted beyond repair		 19:28
[Coke] lizmat: I assumed we'd do something like PS1 for the main and PS2 for the secondary prompt (but allow the same substitutions in both) 21:26
and yes, you can set the env in your startup, or on the command line, etc. 21:27
msiism: I have a POC with \t, \A ( from the bash options) and \i to show the line number. 21:29
anyone have suggestions on a) which of these to support: www.cyberciti.biz/tips/howto-linux...rompt.html and b) if we should follow the same naming convention.
msiism Name of the shell seems to be common as well. 23:18
Bash and OpenBSD’s KSH use `\s` for that.
[Coke]: By line number you mean that result count that is in the default prompt, right? 23:19
Not sure about `\t` and `\A`. I wonder if it would be better to use a single datetime symbol that optionally accepts a format specification. 23:23
And if you have name of the shell, it does make sense to have the shell's release version too. That's `\v` in both Bash and OpenBSD KSH. 23:25
I mean, I'm probably only going to use this whole feature to have less in my prompt, but these are the ones I'd imagine to be quite helpful at times. 23:26
timo fish has it where you provide a function that prints stuff and you can use words instead of escape sequences for a bunch of stuff 23:39
msiism Sounds like Bash's PROMPT_COMMAND. 23:42
timo [...] `set stat (set_color red)"[$last_status]"(set_color normal)` [...] `string join '' -- (set_color green) (prompt_pwd) (set_color normal) $stat '>'` 23:43
msiism The problem with that (as with my suggestion to use a single datetime symbol followed by a spec) is that it's not very convenient for applying quick temporary changes. 23:45
timo what kinds of things do we expect people will find useful in the raku repl prompt? 23:46
msiism The result count, some sort of time display as well as the REPL's name and version are probably sufficient. 23:49
And adding arbitrary literal characters to it. 23:50
msiism Also, the standard `> ` at the end should probably better not be hard-coded. 23:54
timo right 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!