|
00:13
ds7832 joined
00:29
ds7832 left
01:06
hulk joined,
kylese left
02:02
guifa left,
sibl joined
02:15
hulk left,
kylese joined
02:23
sibl left
03:00
annamalai joined
03:16
apogee_ntv left
03:18
apogee_ntv joined
04:38
Aedil joined
06:06
gugod left,
bloatable6__ left,
snonux left,
Voldenet left,
cm left,
lizmat left,
andinus left,
coleman left,
evalable6 left,
gfldex left,
constxd left,
simcop2387 left,
tejr left,
disbot2 left,
bd3i left,
kurahaupo left,
ShimmerFairy left,
kst left,
avuserow left,
dustinm` left,
mtj left,
topnep left,
merp left,
ugexe left,
destroycomputers left,
PotatoGim left,
summerisle left,
human-blip left,
[Coke] left,
perryprog left,
_________ left,
rba left,
SmokeMachine left,
nicole left,
sftp left,
BooK left,
thaewrapt left,
jast left,
JRaspass left
06:10
human-blip joined,
kurahaupo joined,
bloatable6__ joined,
snonux joined,
Voldenet joined,
ShimmerFairy joined,
[Coke] joined,
cm joined,
perryprog joined,
kst joined,
avuserow joined,
dustinm` joined,
lizmat joined,
andinus joined,
mtj joined,
_________ joined,
rba joined,
topnep joined,
coleman joined,
evalable6 joined,
gfldex joined,
constxd joined,
merp joined,
simcop2387 joined,
ugexe joined,
SmokeMachine joined,
nicole joined,
sftp joined,
tejr joined,
BooK joined,
destroycomputers joined,
thaewrapt joined,
PotatoGim joined,
disbot2 joined,
bd3i joined,
summerisle joined,
JRaspass joined,
jast joined
06:11
gugod joined
06:13
abraxxa joined
06:19
abraxxa left,
abraxxa joined
|
|||
| disbot2 | <librasteve> SmokeMachine: Selkie::UI looks very nice. How are “routines” like App { … } and so on implemented… is it maybe a raku subroutine being passed a block? | 07:05 | |
| <librasteve> weekly: github.com/FCO/Selkie-UI/tree/main/examples | 07:06 | ||
| notable6 | librasteve, Noted! (weekly) | ||
| disbot2 | <librasteve> weekly: usercontent.irccloud-cdn.com/file/....35.36.gif | 07:07 | |
| notable6 | librasteve, Noted! (weekly) | ||
| SmokeMachine | librasteve: exactly… github.com/FCO/Selkie-UI/blob/72f3...kumod#L137 | 07:11 | |
| disbot2 | <librasteve> looks very nice … seems there is a fine line in raku between “pseudo DSL” … such as this, Air::Functional, Cro routes (by which I mean making a library that populates a lot of domain specific subroutines into standard Raku) and a “full DSL” with Grammar, Action and Slangify. | 07:18 | |
|
07:20
Sgeo left
07:22
peder left
|
|||
| SmokeMachine | I think there is a 3rd intermediate one, something like this: github.com/m-doughty/MCP-Server/bl...erver.raku or Red that uses “new kind of classes” | 07:26 | |
|
07:26
peder joined
|
|||
| SmokeMachine | That “subvertes” classes, comments, etc… | 07:28 | |
|
07:44
eseyman left
07:45
manu_ joined
07:46
manu_ is now known as eseyman
|
|||
| SmokeMachine | librasteve: 👆 | 08:30 | |
| disbot2 | <librasteve> yeah - very good point | 09:04 | |
| SmokeMachine | librasteve: I'm thinking on playing with this idea: usercontent.irccloud-cdn.com/file/....55.03.gif | 09:58 | |
|
09:59
dano left
|
|||
| SmokeMachine | librasteve: an look how simple the code is: www.irccloud.com/pastebin/Dsx4swqV/zelfie.raku | 10:00 | |
| lizmat | nice | 10:08 | |
| my @items = %list.keys.sort.grep: *.so; # wouldn't it make sense to first grep and then sort ? | |||
| SmokeMachine | Yes, it would | ||
| disbot2 | <librasteve> that is lovely - ship it! | 10:17 | |
| <librasteve> what do you think of a DSL editor - is it too hard? (I heard that it may need new Selkie edit features) | 10:19 | ||
|
10:20
librasteve_ joined
10:27
guifa joined
|
|||
| SmokeMachine | librasteve: it will be shipped inside Selkie::UI… as a bin/ | 10:31 | |
| librasteve_: about the DSL editor, I still don’t get how it would/do. Could you explain a bit more about that, please? | 10:32 | ||
|
10:47
dano joined
|
|||
| disbot2 | <librasteve> something like this langium.org/showcase/arithmetics/ - It would take a mydsl.rakumod with Grammar and Actions and a heredoc for DSL source on the edit side (syntax hilites can wait for v2.0) where the preview pane is text made by the actions (eg just say ~$match.made) (other view filters such as pdf or html can wait for v3.0) | 10:48 | |
| <librasteve> even better would be a 3rd (openable) pane for Grammar::Tracer output | 10:50 | ||
|
10:50
guifa left
10:52
guifa joined
10:56
guifa left
10:58
Guest6550 joined,
Guest6550 left
10:59
Guest6550 joined,
guifa joined
|
|||
| Guest6550 | I have been playing with raku for a bit. There is one thing I have a problem with, though raku is a pretty awesome language. I have this problem with regex where I want to match a string, but part of it I would need to negate as in this should not appear in the text I have no idea how to go about searching for this in the documentation in the first place | 11:01 | |
|
11:02
oodani left
11:03
oodani joined
11:04
guifa left
11:06
oodani left
11:07
oodani joined
|
|||
| SmokeMachine | librasteve_: something like what I wrote for grammars? | 11:11 | |
| lizmat | Guest6550: could you give an example of a string that should match, and one that shouldn't? | 11:13 | |
| Guest6550 | lizmat something like Randomtext.S01 should not be a match for the reason it contains S\d\d and Random.other.Text.2026 should be a match since it does not contain the S\d\d | 11:15 | |
| SmokeMachine | librasteve_: I plan to continue playing with that, but for now this is how it looks: usercontent.irccloud-cdn.com/file/....13.18.gif | ||
| but I plan to make the output prettier... I also think we could color each part of the output with the same color we color each token and the part it matches on the string... but I don't know how to add colors to editable texts... | 11:17 | ||
| lizmat | Guest6550: well, one way to do that is: !$string.contains(/ S\d\d /) so simply negating the result of .contains | 11:18 | |
| Guest6550 | lizmat: that seems much simpler than the regex solution I was trying to build :) | 11:19 | |
| lizmat | it's still a regex solution of sorts :-) | 11:20 | |
| .contains is just a handy shortcut if you just want to know *if* something is in a string, not necessarily what or where | 11:21 | ||
| and by negating that, if something is *not* in a string :-) | 11:22 | ||
| Guest6550 | yes, it does seem to be much more managable than building a more complex regex :-) | ||
|
11:37
Voldenet left,
Voldenet_ joined,
Voldenet_ is now known as Voldenet
11:38
[Coke] left
12:05
[Coke] joined
|
|||
| disbot2 | <librasteve> oh sorry I had forgotten that you already made selkie-ui-grammar-playground - fantastic (I blame days of working on byzantine FatRatStr rounding code) | 12:41 | |
| <librasteve> brilliant - I hope to get DSL back to the foreground soon | 12:42 | ||
|
12:58
eseyman left
12:59
manu_ joined,
manu_ is now known as eseyman
13:16
lizmat left
13:32
shareable6 left
13:34
shareable6 joined
13:37
lizmat joined
13:47
lizmat left
13:48
lizmat joined
14:09
tejr left
14:10
Aedil left
14:11
tejr joined
14:18
Aedil joined
|
|||
| disbot2 | <melezhik.> news.ycombinator.com/item?id=48037017 | 14:57 | |
| <melezhik.> Please vote if you like it , thanks | 14:58 | ||
| [Coke] | are there any http user agents in the ecosystem that respect 429 and will retry? | 15:26 | |
| (Doesn't look like HTTP::UserAgent does) | |||
| (nor Cro) | 15:28 | ||
| coleman | automatic retry is tricky, because you definitely need to know about the context of the application to make a good decision. in Go there's some libraries that wrap the standard client, but I always want to roll my own because HTTP is so low level, and I always need to do some funky config | 15:38 | |
| but Go isn't as expressive, so a wrapper library is probably going to be more powerful in Raku | |||
| [Coke] | I would definitely want guardrails on it (on x number of retrys, what's the algorithm for waiting, etc.) | 15:39 | |
| In the meantime, will write my own for this module that is using HTTP:UA and wrap all the http calls in it rather than doing a subclass of the UA | 15:40 | ||
| coleman | here is some inspiration for the kinds of considerations www.envoyproxy.io/docs/envoy/lates...etrypolicy | 15:42 | |
| envoy has all the complicated stuff laid out in its equally complicated configuration | 15:43 | ||
| [Coke] | nice, thanks. | 15:45 | |
| ugexe | `RAKUDO_RAKUAST=1 make install && RAKUDO_RAKUAST=1 make test` now pass. There is a new CI step that runs those commands to prevent it from regressing | 16:36 | |
| one improvement would be related to dos. for instance any todo that is currently fixed in rakuast could still regress since the test is still marked todo. however, we can't just unmark all todos that are passing in rakuast because they may be failing on the legacy build | 16:38 | ||
| [Coke] | could have the CI for rakuast change the test files before running them. | 16:41 | |
| lizmat | fwiw, for these purposes I'm more thinking of a module that would allow doing spectests with various types of settings, RAKUDO_RAKUAST=1 just being one of them | 16:44 | |
| ugexe | having such a thing requires a bit of nuance. for example we don't encode what frontend is being used in e.g. precomp hashes so it is easy to end up with precomp files that wont work if you're just setting and unsetting RAKUDO_RAKUAST=1 without cleaning up all the stale artifacts | 16:46 | |
| lizmat | cleaning up stale artefacts would be part of such a module | ||
| ugexe | well they would only be stale in the context of whatever test run is using that module. they aren't neccesarily stale from the perspective of the user | 16:47 | |
| in other words such a tool should probably be for internal use only | 16:48 | ||
| lizmat | true, but this module wouldn't be run by most used, only hard core core devs | ||
| *users | |||
| ugexe | but when it doesn't get run by a hard core dev how much time will someone have to spend figuring out the issue? | ||
| lizmat | e.g. a RAKU_TEST_DIE_ON_FAIL=1 at make spectest level would be nice | ||
| quitting at the first test that fails | 16:49 | ||
| ugexe | the errors are not at all intuitive, and actually are obnoxiously good at giving errors that are red herrings | ||
| i wonder if just encoding the front end into the compiler id would fix a lot of that | 16:52 | ||
| it looks like more spectests are failing now than last time i checked | 17:11 | ||
| lizmat | could be some sprintf fallout | 17:13 | |
| I just realized that the 6.e sprintf tests are not being run in "make spectest" :-( | |||
| looking at that now | 17:14 | ||
| ugexe | t/spec/S06-currying/assuming-and-mmd.t and other tests in S06-currying | ||
| lizmat | ah? | ||
| passes for me with RakuAST ? | 17:15 | ||
| the all pass? | 17:16 | ||
| ugexe | did you build rakudo with rakudo_rakuast=1 after cleaning up the stale artifacts? | ||
| lizmat | nope, just running with RAKUDO_RAKUAST after cleaning up t/spec/packages/Test-Helpers/.precomp | 17:18 | |
| lizmat rebuilds | |||
| ugexe | you would need to make clean and then rebuild rakudo with the rakuast frontend | ||
| then that test will fail regardless of if you run it with rakudo_rakuast=1 | 17:19 | ||
| lizmat | This element has not been resolved. Type: RakuAST::Var::Lexical ?? | 17:22 | |
| ugexe | i dont know what the context of that error is but its probably some other stale artifacts | 17:24 | |
| lizmat | looks like all tests using "is-primed-call" fail | 17:25 | |
| ugexe | to be clear i might have broken said tests as well. i dont have time to investigate right now though | 17:26 | |
| lizmat | ah, maybe not | ||
|
17:27
human-blip left
17:29
human-blip joined
|
|||
| lizmat | looks like all the tests using &anonslurp fail | 17:30 | |
| my head is just getting into sprintf formats again, so I'm going to focus on that for now... | 17:31 | ||
|
17:38
ds7832 joined
17:57
ds7832 left
19:31
abraxxa left
19:33
abraxxa joined
19:39
abraxxa left
19:57
annamalai left
20:36
guifa joined
20:49
lizmat left,
lizmat_ joined
20:51
lizmat_ left,
lizmat joined
21:13
guifa left
|
|||
| [Coke] discovers that github.com/raku-community-modules/.../issues/77 is still open. :P | 22:47 | ||
| ah, it's github.com/Raku/whateverable/issues/394, really. | 22:51 | ||
| is REA indexed by raku.land? | 23:09 | ||
|
23:29
Sgeo joined
23:39
librasteve_ left
23:52
Guest6550 left
|
|||
