|
01:02
kjp left
01:03
kjp joined
01:04
kjp left
01:05
kjp joined
01:35
CultOfTheGlenda joined
|
|||
| CultOfTheGlenda | Good morning. I probably have another stupid question. I want to add some sort of operator or language construct. It will look something like | 01:39 | |
| do-when $t { | |||
| #some callable here | |||
| } | |||
| In this case $t is a DateTime. I can achieve what I want by using a subroutine or prefix operator for the most part but I have to use a comma after the $t. Essentially I want that gone. | |||
| I think I've read that things like for loops are written in raku in an operator or grammar. That is basically what I am looking to do. If someone could point me to that, that would be great. Just trying to learn more of the deeper innerworkings. Thanks guys. | 01:46 | ||
| Voldenet | I think there is such operator already | 01:50 | |
| m: my $x = DateTime.now; with $x { .say }; # defined | 01:51 | ||
| camelia | 2025-09-25T03:51:25.061270+02:00 | ||
| Voldenet | m: my DateTime $x;; with $x { .say }; # not defined | ||
| camelia | ( no output ) | ||
| Voldenet | s/operator/keyword/ | ||
| there are plenty of ways to do exactly what you want or something similar - the easiest (but not precisely what you want) is a sub | 01:53 | ||
| m: sub do-when ($y, &x) { with $y { x($y) } }; do-when DateTime.now, { .say } # no hacking, boring code | |||
| camelia | 2025-09-25T03:54:00.063156+02:00 | 01:54 | |
| Voldenet | it requires `,` after the parameter, so it's not as good | ||
| CultOfTheGlenda | Yeah, this is an exercise to learn more about raku(do). What's the hard way? | 01:56 | |
| Voldenet | other option is a slang (where you add subrules to a grammar) - it's sufficiently complex at the moment because of QAST/RakuAST differences | 01:57 | |
| github.com/0racle/raku-Slang-Other...ree/master example slang, but you can find more in raku.lang | |||
| raku.land* | 01:58 | ||
| CultOfTheGlenda | Thanks, I was afraid slang was the way. That comma is looking better already. | 01:59 | |
| Voldenet | I think the most modern way would be rakuast macros | ||
| there's also keywords on the HOW levels, but I think they're limited to defining class-like keywords, maybe I'm not seeing bigger picture though | 02:00 | ||
| example of that approach (probably not applicable to defining entirely new syntax) github.com/raku-community-modules/.../tree/main | 02:01 | ||
| CultOfTheGlenda | One thing that confuses me about raku is arrays in a hashmap. I will then point another array at that nested array thusly | ||
| my @ss := %hashmap<key-that-points-to-array>; | |||
| When I do that @ss becomes a nested array. Why is that? | |||
| I've been doing raku for years now, and that still trips me up. | 02:02 | ||
| Voldenet | m: my %x = a => [1]; say %x<a>; my @n := %x<a>; say @n | 02:03 | |
| camelia | [1] [1] |
||
| Voldenet | you mean the above? | ||
| m: my %x = a => [1]; say %x<a>; my @n := %x<a>; say @n; my @nested = %x<a>; | 02:04 | ||
| camelia | [1] [1] |
||
| Voldenet | m: my %x = a => [1]; say %x<a>; my @n := %x<a>; say @n; my @nested = %x<a>; say @nested | ||
| camelia | [1] [1] [[1]] |
||
| Voldenet | the last one copies the scalar that contains an array | 02:07 | |
| you can see incarnations of this in many places | 02:08 | ||
| m: my %x = a => [1, 2, 3]; for %x<a> { .say } | |||
| camelia | [1 2 3] | ||
| .landyacht. | m: my %x = a => [1, 2, 3]; for @%x<a> { .say } | 02:09 | |
| evalable6 | (exit code 1) Type List does not support associative indexing. in block <unit> at /tmp/Szu2Abr8kU line 1 |
||
| .landyacht. | ay? | ||
| m: my %x = a => [1, 2, 3]; for @(%x<a>) { .say } | |||
| evalable6 | 1 2 3 |
||
| Voldenet | The above is what I used to do, but | ||
| .landyacht. | eesh, don't love that... | ||
| Voldenet | it's not quite right | ||
| it calls .list on that and COPIES the contents | |||
| but what you can do is say "I don't want the container but the contents" | 02:10 | ||
| m: my %x = a => [1, 2, 3]; for %x<a><> { .say } | |||
| camelia | 1 2 3 |
||
| .landyacht. | I'm sure I remember this working this way to make something else DWIM nicely, but I can't for the life of me remember what that thing is | ||
| Voldenet | m: my %x = a => [1, 2, 3]; for %x<a>.list { .say } # this is doing the same as @(…) version | 02:13 | |
| camelia | 1 2 3 |
||
| Voldenet | lists aren't expensive, but at the moment I think they're allocated, so they're not as free as the version with <> | 02:16 | |
| there's also Slip version of it, but I don't know how that internally works | 02:24 | ||
| m: my %x = a => [1, 2, 3]; for |%x<a> { .say } # the same as %x<a>.Slip | 02:25 | ||
| camelia | 1 2 3 |
||
| Voldenet | Slip also probably copies the list, that's what benchmarks suggest at the moment: 0x0.st/KmV7.raku | 02:36 | |
|
02:43
greaser|q left,
greaser|q joined,
greaser|q is now known as GreaseMonkey
|
|||
| cultoftheglenda | Back on discord. Thanks for all the information guys. | 03:26 | |
| CultOfTheGlenda | What would the performance impact be of a shlang? | 03:59 | |
|
04:07
CultOfTheGlenda left
|
|||
| Voldenet | CultOfTheGlenda: altering the grammar takes up to 1s, but after that there's hardly any impact | 04:11 | |
| tellable6 | Voldenet, I'll pass your message to CultOfTheGlenda | ||
|
04:34
derpydoo joined
05:24
Aedil joined,
derpydoo left
05:34
Sgeo left
05:50
melezhik joined
06:06
abraxxa joined
06:31
derpydoo joined
07:20
melezhik_ joined
|
|||
| melezhik_ | .tell ds7832: - please take a look at Sparky - looks good for your tasks IMHO - gist.github.com/melezhik/d1f42e5ef...ef5da5b879 | 07:27 | |
| tellable6 | melezhik_, I'll pass your message to ds7832 | ||
|
07:27
melezhik_ left
|
|||