<< 01-04-2025 >>
00:14:30*xet7 quit (Quit: Leaving)
00:16:16*xet7 joined #nim
01:51:07*andreas joined #nim
01:51:07*andreas_ quit (Read error: Connection reset by peer)
02:02:59FromDiscord<morgan (ping with reply)> In reply to @Elegantbeef "Could always just ask": it’s been two years, last update was in a year, there were two issues open for months about the problem
02:03:53FromDiscord<morgan (ping with reply)> if it was a month or few sure yea try to get the original dev to update that
02:04:43FromDiscord<morgan (ping with reply)> and the issue was from him tagging his versions, keeping it defaulting to an older version. no tags at all would’ve been better here actually
02:05:39FromDiscord<morgan (ping with reply)> two years since the last tag, to be specific, a year since the last commit
02:06:59FromDiscord<Elegantbeef> Not implying anything but this is how supply chain attacks start 😄
02:16:19*rockcavera quit (Remote host closed the connection)
02:16:26FromDiscord<zumi.dxy> speaking of tagging↵how is specifying an exact commit "not specific enough" for nimble
02:17:34FromDiscord<Elegantbeef> What do you mean?
02:17:41FromDiscord<zumi.dxy> though I can see how it introduces problems for dependency resolution
02:17:49FromDiscord<zumi.dxy> because it wants like↵version ranges
02:18:13FromDiscord<zumi.dxy> can't really infer anything from a commit hash
02:18:39FromDiscord<Elegantbeef> You can use commits
02:20:14FromDiscord<zumi.dxy> commit ranges?
02:20:35FromDiscord<zumi.dxy> does it defer to vcs to determine "recentness"
02:44:24FromDiscord<Elegantbeef> Using git hashes seems ill advised for humans
02:55:08FromDiscord<morgan (ping with reply)> In reply to @Elegantbeef "Not implying anything but": that is a fair point
02:55:36FromDiscord<morgan (ping with reply)> @treeform if you see this, please add a new tag on latest on chroma
02:55:39FromDiscord<morgan (ping with reply)> there
03:40:10FromDiscord<treeform> Wait why?
03:49:00FromDiscord<Elegantbeef> Cause your latest release is almost 6 months before a change in code
04:27:21FromDiscord<litlighilit> How can I get the parent type (super class) of a object in Nim?
04:39:45FromDiscord<graveflo> sent a code paste, see https://paste.rs/Jchpw
04:40:28FromDiscord<graveflo> sent a code paste, see https://paste.rs/xqI2e
04:40:29FromDiscord<Elegantbeef> https://github.com/beef331/oopsie/blob/master/src/oopsie.nim has `super` and `rootSuper`
04:42:47*SchweinDeBurg quit (Quit: WeeChat 4.7.0-dev)
04:42:59FromDiscord<graveflo> sent a code paste, see https://paste.rs/WYCKX
04:42:59FromDiscord<morgan (ping with reply)> In reply to @treeform "Wait why?": the latest release is from 2.5 years ago, not long after a significant bug in oklab to rgb conversion was fixed, and it's still causing problems. there's two issues from almost a year ago about this, first about the bug and second that it was fixed but requires a workaround to use the fix. just bumping the version with a new tag and a new nimble version would fix this
04:43:13*SchweinDeBurg joined #nim
04:43:38FromDiscord<Elegantbeef> Right grave the var is lifted
04:44:22FromDiscord<graveflo> meaning?
04:46:15FromDiscord<Elegantbeef> meaning `var` is no longer allocated on the stack and lives inside the closure environment
04:48:25FromDiscord<graveflo> I would really like code such as the snippet I posted to compile and it seems fine to me. It's too difficult for the compiler to track, probably because it has to hand the closure to the dispatcher or something. I also don't know how multiple async back-end work and if the transformations are the same. As in, would this always be safe regardless of async backend
04:49:22FromDiscord<Elegantbeef> Well it cannot cause `var T` cannot be captured and enabling it requires borrowing semantics to ensure the lifetime
04:49:29FromDiscord<Elegantbeef> Until then use `ptr T` insteaad of `var T`
04:50:41FromDiscord<graveflo> yea that is what I was going to do. trying to write a macro that will transform async code to blocking and I know about `multisync` but I don't like the way it works. The checking should just be simple as the second closure must be awaited in the first one. Otherwise the future can run off on it's own
04:51:43FromDiscord<graveflo> ah idk... the macro is kind of hard to write. I think I'm going to have to deal with both typed and untyped ast at the same time and I don't even know if that will screw things up. I have to type the `proc` body to know when to switch from pointer to var but I have to emit from untyped ast
05:01:25FromDiscord<graveflo> sent a code paste, see https://paste.rs/HTG11
05:02:02FromDiscord<graveflo> because the NimNode is the type I presume
05:02:34FromDiscord<leorize> you're not stacking your macros correctly
05:03:05FromDiscord<leorize> `toTyped(x)` will actually run on `Symbol x`, not the input
05:03:30FromDiscord<leorize> the input must enter the macro in a `typed` context
05:03:54FromDiscord<graveflo> right that makes sense. I'm trying to get both though
05:04:09FromDiscord<leorize> emit a call to your other macro
05:04:23FromDiscord<graveflo> ah that makes so much sense
05:04:38FromDiscord<leorize> `StmtList(original, Call(processTyped, original))` \<- something like that
05:04:55FromDiscord<graveflo> wait that might not work either though because I need the typed ast before the untyped macro finishes.. I guess I need 3 stages then?
05:05:27FromDiscord<leorize> that could work
05:05:37FromDiscord<leorize> but know that typed ast is very unstable
05:06:07FromDiscord<leorize> a simple `macro toTyped(x: typed): untyped = x` can fail simply because the input passed through a macro
05:06:10FromDiscord<graveflo> yea that's one of the issues. I would start with typed ast and then emit it as if it were untyped but I know better then that at this point
05:06:59FromDiscord<leorize> if you need typed ast workarounds, look at cps' `normalizedast` and `rewrites`
05:07:21FromDiscord<graveflo> last I asked a question about that kind of stuff I was told to look at cps source code and I understood that was probably not the way to go
05:07:21FromDiscord<leorize> though I'll say that performance will plummet as a result
05:07:29FromDiscord<graveflo> yea there it is
05:07:53FromDiscord<graveflo> not their fault but that stuff is way too involved
05:08:46FromDiscord<leorize> every workaround in there is found through blood and tears
05:09:05FromDiscord<leorize> so if you're ever stuck it's a good reference on how to untangle your issue
05:09:17*ntat joined #nim
05:53:42FromDiscord<user2m> Is there a way to capture everything thats written to the terminal ?
05:54:09FromDiscord<graveflo> input or output? posix or no?
05:56:48FromDiscord<user2m> Output and I'm on windows so no posix
05:56:51FromDiscord<leorize> if it's written by your program, then override stdin/out/err
05:57:05FromDiscord<leorize> if it's written by programs you run, then osproc already got you covered
06:02:57FromDiscord<Phil> Out of curiosity, anyone have roughly an idea how much effort implementing SSE would be? ↵Been wondering if I want to play around with it but I don't think prologue supports it atm
06:17:33*derpydoo joined #nim
06:21:59*derpydoo quit (Ping timeout: 244 seconds)
06:22:44*derpydoo joined #nim
06:36:39*nils` quit (Ping timeout: 245 seconds)
06:42:42FromDiscord<nitely_> In reply to @isofruit "Out of curiosity, anyone": I'm about to for hyperx, but AFAIK, any server that supports streaming supports SSE
07:25:50*andreas quit (Quit: Konversation terminated!)
07:26:02*andreas joined #nim
07:28:09FromDiscord<morgan (ping with reply)> afaik any normal computer cpu that's vaguely not ancient supports sse of some sort
07:32:17FromDiscord<Elegantbeef> Wrong SSE↵(@morgan (ping with reply))
07:32:30FromDiscord<Elegantbeef> Phil is talking webdev not SIMD
07:32:45*nils` joined #nim
07:34:14FromDiscord<morgan (ping with reply)> oh lol
07:34:21FromDiscord<morgan (ping with reply)> what's sse in that context?
07:49:20Amun-Rahttps://en.wikipedia.org/wiki/Server-sent_events
07:56:04*nils` quit (Ping timeout: 260 seconds)
08:00:09*ntat quit (Ping timeout: 260 seconds)
08:11:33FromDiscord<morgan (ping with reply)> ah
08:14:19Amun-Rathere are way too many acronyms in IT
08:27:39FromDiscord<lainlaylie> initialisms too
08:32:00*Zevv joined #nim
08:32:26Zevvwho's in charge of the forum these days, is that still PMunch?
08:35:54FromDiscord<Phil> In reply to @morganalyssa "ah": Basically: Half a websocket connection that maybe is easier to manage
08:46:27FromDiscord<nnsee> In reply to @Zevv "who's in charge of": yes
08:51:58Zevvok thank you
09:06:03*nils` joined #nim
10:00:51*m5zs7k quit (Ping timeout: 276 seconds)
10:20:29*m5zs7k joined #nim
10:46:10*Guest50 joined #nim
10:46:26*Guest50 quit (Client Quit)
11:46:42FromDiscord<enthus1ast.> We need Webserver streaming support in general in nim Land.↵In both directions.
11:47:29FromDiscord<enthus1ast.> The current situation is kinda sad imho 🙂
11:48:27FromDiscord<enthus1ast.> The only server I think has streaming support is guildenstern(?)
11:51:25FromDiscord<enthus1ast.> I wish mummy would have streaming support, or could grand direct access to the socket, but I can imagine that it's not so easy to add
11:51:36FromDiscord<enthus1ast.> (edit) "grand" => "grant"
11:58:56*derpydoo quit (Quit: derpydoo)
12:11:20FromDiscord<zumi.dxy> In reply to @enthus1ast. "The only server I": using it rn↵and I can say↵it has motivated me to build my own routing, handling, etc. etc.↵it is very spartan
12:11:49FromDiscord<zumi.dxy> only reason I wanted streaming support is I don't want 8GB uploads to kill the server lmao
12:12:11FromDiscord<enthus1ast.> Yes that's the reason I let nginx handle uploads
12:13:01FromDiscord<enthus1ast.> Then when done I move it to the final destination
12:13:54FromDiscord<zumi.dxy> do you have like a separate endpoint just for uploads
12:14:23FromDiscord<zumi.dxy> e.g. a file upload form where its `action` is somewhere else
12:52:33FromDiscord<arnetheduck> In reply to @enthus1ast. "We need Webserver streaming": chronos can stream
12:53:07FromDiscord<arnetheduck> as can nim-websock, from what I remember
12:55:49FromDiscord<arnetheduck> chronos also supports SSE / chunking: https://github.com/status-im/nim-chronos/blob/b55e2816eb45f698ddaca8d8473e401502562db2/chronos/apps/http/httpserver.nim#L1546
13:10:41*ntat joined #nim
13:38:27FromDiscord<enthus1ast.> Ah this is good to know. I never took the time to have a closer look at chronos despite i always wanted to
13:43:36Amun-RaI have an enum with associated strings, is there a way of getting enum name in this case?
14:50:31*ntat quit (Quit: leaving)
14:56:19*ntat joined #nim
15:34:18FromDiscord<arnetheduck> In reply to @enthus1ast. "Ah this is good": well, it doesn't do anything fancy really, it's main feature is that it ... just works
15:34:27FromDiscord<arnetheduck> (edit) "In reply to @enthus1ast. "Ah this is good": well, it doesn't do anything fancy really, it's main feature is that it ... just works ... " added "(well)"
15:34:34FromDiscord<arnetheduck> (edit) "works (well)" => "works..."
16:07:41FromDiscord<alehander92> is there a reason
16:07:57FromDiscord<alehander92> httpclient `add` API-s can't work with streams
16:09:09FromDiscord<alehander92> i assume there wasn't such a common usecase, but it seems possible to augment the code so it accepts streams (as it already can create itself a stream from a file, just without letting someone pass a custom stream)
17:05:16*beholders_eye joined #nim
17:22:30*ntat quit (Quit: leaving)
17:33:18*ntat joined #nim
18:03:37*beholders_eye quit (Ping timeout: 244 seconds)
18:20:46*tk is now known as a123456789012345
18:20:55*a123456789012345 is now known as tk
18:59:32*derpydoo joined #nim
20:09:58FromDiscord<threefour> What's the best way to do a static `char` to `int` lookup table? Not every `char` value is used, so there would be holes.
20:11:35FromDiscord<treeform> HashMap[char, int] if you want to be like python, array[int, int] if you want to be like C. cleaver bit fiddling formula if you want to be cool like Nim.
20:11:45FromDiscord<Elegantbeef> `array[char, int]` is the quickest but uses `256 int`
20:12:03FromDiscord<Elegantbeef> `Table[char, int]` uses less memory but has a hash overhead and is not a direct index operation
20:14:15FromDiscord<treeform> In reply to @morganalyssa "the latest release is": Done: https://github.com/treeform/chroma/releases/tag/v1.0.0 thank you for brining this to my attention.
20:14:42FromDiscord<treeform> Is this still an issue for you? https://github.com/treeform/chroma/issues/41
20:15:35FromDiscord<morgan (ping with reply)> awesome
20:16:11FromDiscord<morgan (ping with reply)> i will try it out but bumping tag and nimble version was all i did before in my fork to make it work
20:17:58FromDiscord<morgan (ping with reply)> yep looks fine
20:18:48FromDiscord<morgan (ping with reply)> and 41 is fixed with this
20:26:27FromDiscord<treeform> Great. If you see any other issues like this. Feel free to ping.
21:07:35*derpydoo quit (Quit: derpydoo)
21:08:33*ntat quit (Quit: leaving)
21:16:55FromDiscord<threefour> In reply to @Elegantbeef "`array[char, int]` is the": This is what I'm attempting, but I get "invalid order in array constructor" presumably because of holes
21:17:14FromDiscord<threefour> I'd like to not have to spam a bunch of useless entries
21:21:29FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/Adbde
21:22:31FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/J0HdM
21:40:08FromDiscord<threefour> I suppose. I was just hoping there was table constructor syntax sugar I was missing.
21:40:43FromDiscord<Elegantbeef> Could make a .... macro
21:41:37FromDiscord<Elegantbeef> Actually don't need a macro
21:43:30FromDiscord<Elegantbeef> We're cooking with bacon
21:43:30FromDiscord<Elegantbeef> sent a code paste, see https://paste.rs/KmHye
21:48:38FromDiscord<threefour> In reply to @Elegantbeef "Could make a ....": It'd be overkill because I just need this singular lookup table. Might as well do it manually.
21:55:53*rockcavera joined #nim
21:57:52FromDiscord<threefour> sent a code paste, see https://paste.rs/jnaQS
21:59:03FromDiscord<dawidek.2137> If you just want less verbose code you could also define a local template that assigns a value to a key in the array and nothing else
21:59:22FromDiscord<dawidek.2137> Dumb but I do it sometimes
21:59:23FromDiscord<Elegantbeef> I feel like the above is pretty terse
23:14:39*zgasma quit (Ping timeout: 244 seconds)