codesections will do 00:00
00:09 sena_kun left
codesections timotimo: wait, how do I add a file to core.d? I thought I could follow the model of the existing files, which seemed to be just adding a couple of lines to the Makefile 00:33
but that didn't seem to do the trick. Do you happen to know what else I need to do?
timotimo i haven't tried it yet myself 00:40
codesections Hmm, ok I'll keep experimenting thanks 00:41
timotimo i would recommend looking at the git log for some of the files in that folder with the -u option to get the diffs 00:47
like git log -u -- src/core.d/one-of-the-files.rakumod
though the renaming of pm6 to rakumod could be trouble 00:48
codesections good idea, thanks
01:40 nebuchadnezzar left
Geth rakudo: codesections++ created pull request #3903:
Revise deprecation warnings to show only on v6.d		 02:36
06:00 domidumont joined 06:01 nebuchadnezzar joined 06:03 vrurg_ joined 06:05 vrurg left 06:15 sjn__ is now known as sjn
Geth rakudo: usev6++ created pull request #3904:
Start to use nqp::execname on the JVM backend		 06:34
07:34 MasterDuke joined 08:01 yuplushi left 08:09 sena_kun joined 08:41 domidumont left 08:52 yuplushi joined 09:06 MasterDuke left 09:25 MasterDuke joined 09:58 yuplushi left
lizmat Files=1336, Tests=113586, 221 wallclock secs (29.67 usr 7.85 sys + 3121.60 cusr 270.75 csys = 3429.87 CPU) 10:34
10:56 yuplushi joined 11:01 MasterDuke left
Geth rakudo/IO-Notification-Change-ValueObjAt: 17f352f510 | (Elizabeth Mattijsen)++ | src/core.c/IO/Notification.pm6
Make IO::Notification::Change a value type

Which should make it better to filter for it, e.g. with Supply.unique(expires => X).		 11:09
rakudo: lizmat++ created pull request #3905:
Make IO::Notification::Change a value type
11:12 finsternis left 11:13 finsternis joined 11:28 MasterDuke joined 11:45 Altai-man joined 11:47 sena_kun left 12:08 [TuxCM] left 12:15 [TuxCM] joined
MasterDuke i think we have a sort of Shlemiel the painter's algorithm situation going on with the whatever slice and reify-at-least 12:44
if i dd $!reified (using the char range example in this case) i get output like: (), (), ("a"), (), ("a"), ("a", "b"), (), ("a"), ("a", "b"), ("a", "b", "c"), (), ("a"), ("a", "b"), ("a", "b", "c"), ("a", "b", "c", "d") 12:48
i.e., it seems to keep recreating it one longer each time, instead of just appending to the existing 12:49
timotimo what on earth? :) :)
could you paste that including tracebacks at every output point? 12:50
MasterDuke add something like `Backtrace.new.Str`? 12:52
timotimo i think so. perhaps .full
MasterDuke hm. that causes a hang in stage optimize. but the original dd didn't cause any output there... 12:57
oh. do i need to move Backtrace.pm6 in core_sources? 12:58
nope. same thing 13:01
13:01 Kaeipi left
MasterDuke what's another way to accomplish this? catch a die? 13:03
13:05 Kaiepi joined 13:26 domidumont joined 13:32 Geth left 13:33 Geth joined
timotimo oh! 13:50
of course 13:51
in order to print out a stack trace
guess what it has to do
iterate over the stack frames :)
13:56 lucasb joined
MasterDuke been distracted. any other way to do it? 14:08
timotimo hmm
put nqp::sin_n(0e0) and breakpoint in gdb
the MVM_dump_backtrace function doesn't do raku-level iteration
MasterDuke ugh, going to be so many... 14:09
timotimo how do you mean? 14:12
you know you can make gdb do it automatically, including resuming :)
commands
call MVM_dump_backtrace(tc)
c
end
MasterDuke oh, right. forgot about that
timotimo i need it so often ... :D 14:13
MasterDuke huh. `b sin` doesn't break 14:14
timotimo yeah, you'll want to breakpoint that in interp.c 14:18
sin would also not have tc available
MasterDuke got a 3.5mb log file 14:23
33k lines
partial output
timotimo it'll probably be extremely duplicatious 14:24
MasterDuke gist.github.com/MasterDuke17/81ba7...746d122ae2 some of the output 14:26
timotimo i kind of feel like just looking at the code near those lines should give us a good idea of what's going on 14:30
it does seem like what you said
like it's re-making the range into a list every time it grabs a new value
and it goes via AT-POS, maybe that's the wrong? 14:31
MasterDuke github.com/rakudo/rakudo/blob/mast...st.pm6#L40 14:32
looks like 4 reify-at-least for every AT-POS 14:34
timotimo aw dang, the tampermonkey script is still confused by the core.c lines and thinks they're moarvm code 14:42
MasterDuke easy to just hardcode them? 14:44
timotimo fixed it 14:46
just needed to allow for `.[a-zA-Z]`
and updated the gist with the script in it 14:48
MasterDuke is going to be distracted for a bit, currently have both the three-year-old and the two-month-old by myself
timotimo :)
i kind of hate that in order to see the function names at the end of the lines i have to horizontally scroll in the gist 14:49
MasterDuke hm, tampermonkey doesn't see any updates available
yeah, too bad that box isn't resizeable 14:50
timotimo gist.github.com/timo/7cfe71a667bbd...0431da45a4 - perhaps have to redownload this dunno 14:51
next step: allow the user to embed a little iframe with the source code in it 14:52
maybe easier to generate a vim quicklist so you can use your favourite editor to navigate these backtraces 14:54
MasterDuke updated, timotimo++
15:46 sena_kun joined 15:47 Altai-man left 15:48 squashable6 left 15:49 squashable6 joined 17:16 MasterDuke left 18:15 Xliff joined
Xliff . 18:16
\o
nwc10 o/ 18:21
18:56 domidumont left, domidumont joined 18:57 domidumont left, domidumont joined, domidumont left
Kaiepi blegh 19:04
nvm this is more a moarvm thing 19:05
19:15 MasterDuke joined 19:45 Altai-man joined 19:47 sena_kun left 20:27 literal_ is now known as literal 21:25 pamplemousse joined 21:30 pamplemousse left
MasterDuke that reify-at-least code is old. most recent change is three years ago and the bulk of it is four years old 21:32
21:40 pamplemousse joined 21:44 Xliff left 22:57 leont left 23:13 pamplemousse_ joined 23:17 pamplemousse left 23:20 Altai-man left 23:21 MasterDuke left 23:40 Kaiepi left 23:47 Kaiepi joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!