| 00:00:01 | * | junland quit (Quit: %ZNC Disconnected%) |
| 00:00:09 | Yardanico | most of it is not really useful in 99.9% of issues |
| 00:00:17 | Yardanico | imo |
| 00:00:46 | disruptek | i will say this about concepts: |
| 00:00:57 | disruptek | anything that improves compile-time speed is very welcome. |
| 00:01:09 | FromDiscord | <Varriount> Are concepts slow? |
| 00:01:18 | disruptek | beyond slow. |
| 00:01:41 | disruptek | like, i know this code doesn't work and it'll take minutes for me to find out why. |
| 00:01:47 | * | junland joined #nim |
| 00:02:00 | Yardanico | are you sure they are *that* slow? |
| 00:02:06 | Yardanico | why are there no open issues about that? :P |
| 00:03:21 | disruptek | because an issue will serve no purpose. |
| 00:03:29 | FromDiscord | <Varriount> From what I recall, concepts make heavy use of `compiles()`. |
| 00:03:35 | Yardanico | well yes |
| 00:03:38 | Yardanico | but anyway, I don't think that's a good attitude to have :) |
| 00:03:41 | Yardanico | good night |
| 00:03:46 | disruptek | gn |
| 00:04:05 | disruptek | there's no point to creating an issue for a feature that is slated for redesign/reimpl. |
| 00:04:18 | disruptek | not to merely complain about performance, anyway. |
| 00:04:38 | FromDiscord | <Varriount> Well, it lets others know that the issue is known. |
| 00:04:57 | disruptek | it is known. |
| 00:06:28 | FromDiscord | <Varriount> Yes, but I may not know that, unless someone tells me. If I'm a new user, use concepts, and discover that they are slow, looking at the issue tracker would suggest that the issue isn't known. |
| 00:08:38 | disruptek | anything that improves compile-time speed of concepts is very welcome. |
| 00:10:51 | disruptek | if you don't know, now you know |
| 00:17:38 | * | lritter quit (Quit: Leaving) |
| 00:26:17 | FromDiscord | <exelotl> This one's a stinker https://github.com/nim-lang/Nim/issues/15218 |
| 00:26:19 | disbot | ➥ Const variant object fields are incorrectly set to zero ; snippet at 12https://play.nim-lang.org/#ix=2uXz |
| 00:26:53 | disruptek | oh wow. |
| 00:27:33 | disruptek | nice job. |
| 00:28:07 | FromDiscord | <exelotl> thanks xD |
| 00:30:55 | FromDiscord | <exelotl> The real travesty here is that this bug broke my poor dog's rocket. https://media.discordapp.net/attachments/371759389889003532/746889166825979916/Screenshot_from_2020-08-23_01-29-59.png |
| 00:31:13 | disruptek | ouch. |
| 00:31:43 | disruptek | you hate to see it. |
| 00:32:08 | FromDiscord | <exelotl> ya really do |
| 00:32:12 | FromDiscord | <Elegant Beef> I dont like that your dog's rocket is red |
| 00:33:35 | disruptek | that's such a bad bug. |
| 00:40:05 | * | bung joined #nim |
| 00:46:01 | * | waleee-cl quit (Quit: Connection closed for inactivity) |
| 00:57:22 | FromDiscord | <Varriount> exelotl: Whenever I see your nickname, I always want to draw/create some stylized axelotl icon for you |
| 00:58:08 | FromDiscord | <exelotl> aww |
| 00:58:57 | * | bung quit (Ping timeout: 258 seconds) |
| 00:59:04 | FromDiscord | <exelotl> I'm too attached to my plush lizard tho :( |
| 01:00:09 | * | Jesin joined #nim |
| 01:00:15 | * | Jesin quit (Remote host closed the connection) |
| 01:07:46 | * | Jesin joined #nim |
| 01:07:46 | * | leorize quit (Remote host closed the connection) |
| 01:07:46 | * | Jesin quit (Remote host closed the connection) |
| 01:08:19 | * | leorize joined #nim |
| 01:09:33 | * | Jesin joined #nim |
| 01:10:02 | disruptek | varriount: i think my concepts are actually never going to finish compiling. |
| 01:11:53 | * | thomasross joined #nim |
| 01:12:20 | * | Jesin quit (Remote host closed the connection) |
| 01:17:37 | * | Jesin joined #nim |
| 01:17:58 | disruptek | 1.4hrs is just too long to wait. |
| 01:21:54 | * | Jesin quit (Remote host closed the connection) |
| 01:23:07 | * | aenesidemus_ joined #nim |
| 01:25:54 | leorize | disruptek: have you looked at one of the proposals within the RFC? https://github.com/nim-lang/RFCs/issues/168#issuecomment-593410686 |
| 01:25:55 | disbot | ➥ Concepts and type-checking generics ; snippet at 12https://play.nim-lang.org/#ix=23U8 |
| 01:26:05 | * | aenesidemus quit (Ping timeout: 240 seconds) |
| 01:26:37 | * | Jesin joined #nim |
| 01:26:51 | FromDiscord | <Varriount> disruptek: Despite my complaint about command substitution, I do like using fish. It actually follows the kind of design I was aiming for when writing commandant. |
| 01:27:02 | leorize | that one in particular is interesting to me due to how simple it appears to be |
| 01:27:23 | disruptek | well, i like what i have now. |
| 01:27:28 | leorize | @Varriount `string collect` is a good enough magic :P |
| 01:27:39 | disruptek | i just wish it actually compiled instead of, y'know, just heating up my house. |
| 01:28:10 | disruptek | varriount: you mean the thing about not being about to run a variable? |
| 01:28:24 | disruptek | varriount: you can use `command $foo` to "invoke" foo. |
| 01:28:46 | leorize | I thought @zah said he got some ideas on how to fix the current implementation? |
| 01:28:56 | disruptek | our current concepts are succinct and similar to what gemath suggests, but even simpler. |
| 01:29:07 | FromDiscord | <Hearthstone> > disruptek: Despite my complaint about command substitution, I do like using fish. It actually follows the kind of design I was aiming for when writing commandant.↵Fish as in the fish shell? |
| 01:29:11 | disruptek | it's just that you cannot use negation. like, at all. |
| 01:29:18 | disruptek | hearthstone: yes. |
| 01:29:22 | FromDiscord | <Hearthstone> It's my favourite shell- |
| 01:29:33 | leorize | disruptek: why couldn't you? it's a list of expression |
| 01:29:37 | disruptek | i'm a 3-shell kinda guy. |
| 01:29:56 | disruptek | leorize: i dunno why. |
| 01:30:12 | FromDiscord | <Varriount> @Hearthstone Yes |
| 01:30:41 | FromDiscord | <Hearthstone> Nice |
| 01:30:50 | leorize | fish shell is the best shell |
| 01:30:56 | leorize | well at least until nim shell is a thing :P |
| 01:31:05 | disruptek | leorize: i just pushed a nimph PR that fails spectacularly if you wanna look at concept insanity. |
| 01:31:09 | disruptek | it's not even complex. |
| 01:31:22 | leorize | yes please |
| 01:31:31 | FromDiscord | <Varriount> @Hearthstone The fact that you can't quote command substitutions in Fish is dangerous, which precludes me from using it for scripts to be used in automation, or which accept arbitrary input. |
| 01:31:31 | leorize | if you really want to use concepts, ask @mratsim |
| 01:31:42 | leorize | you can now |
| 01:31:48 | disruptek | no idea what you mean. |
| 01:32:19 | FromDiscord | <Hearthstone> What do you mean? |
| 01:32:21 | leorize | mratsim managed to figure out concepts for many of his projects |
| 01:32:34 | disruptek | i've used them successfully. |
| 01:33:03 | disruptek | https://github.com/disruptek/nimph/pull/140 |
| 01:33:04 | disbot | ➥ nimph 2.0 |
| 01:33:17 | disruptek | there's some shit about nimph-2.0 if anyone is interested. |
| 01:33:29 | FromDiscord | <Clyybber> just looked at it |
| 01:34:23 | disruptek | is it making your balls throb as hard as it is mine? |
| 01:35:48 | FromDiscord | <Clyybber> @Varriount its reported (the slowness of concepts) |
| 01:36:02 | FromDiscord | <Clyybber> disruptek nah |
| 01:36:13 | disruptek | what's wrong with your balls? |
| 01:36:16 | FromDiscord | <Clyybber> not the infinite loop |
| 01:36:27 | FromDiscord | <Clyybber> but its funny |
| 01:36:43 | disruptek | i think it's geometric growth. |
| 01:36:56 | FromDiscord | <Clyybber> i just recentpy started building a pm that uses git submodules |
| 01:37:08 | FromDiscord | <Clyybber> but now that nimph is going that path i'm fine |
| 01:37:51 | leorize | not a big fan of git submodules tbf |
| 01:38:28 | disruptek | one thing i love about submodules is that they show up as links on github. |
| 01:38:33 | FromDiscord | <Clyybber> they can be inconvinient, thats why you build a pm arpund them |
| 01:38:43 | disruptek | so you can jump right to the project at the very commit... it's awesome. |
| 01:38:52 | disruptek | you can see exactly what the dependencies are, etc. |
| 01:38:57 | FromDiscord | <Clyybber> and the big advantage is that people don't need a pm to build your stuff |
| 01:39:03 | FromDiscord | <Clyybber> they are pm independent |
| 01:39:05 | leorize | the only part I hate about them is that they require a git repo |
| 01:39:09 | FromDiscord | <Varriount> @Hearthstone In Bash, `rm "$(some command)"` runs `some command`, captures its output, , and then runs `rm` with the first parameter set to the captured output. |
| 01:39:11 | disruptek | right. like gitnim... "it's just git" |
| 01:39:49 | FromDiscord | <Hearthstone> Yup |
| 01:39:53 | FromDiscord | <Varriount> @Hearthstone If you omit the quotes (`rm $(some command)`) then word splitting is performed. |
| 01:40:02 | FromDiscord | <Hearthstone> Okay |
| 01:40:06 | leorize | @Varriount `rm -- (some command | string collect)` is now a thing :P |
| 01:40:07 | disruptek | i don't have any problem with requiring that every project is in source control. if you don't want that, just don't use nimph. nim is sufficient. |
| 01:40:29 | leorize | disruptek: it just complicates tarball releases |
| 01:40:42 | disruptek | tell me. |
| 01:41:35 | disruptek | clyybber: you know you can use nimph with submodules today. |
| 01:41:52 | leorize | the extra step of pulling all repos bundled in your repo, then remove all `.git` from them before packing into an archive |
| 01:41:59 | leorize | arguably not much of an issue though |
| 01:42:02 | disruptek | yawn. |
| 01:42:05 | FromDiscord | <Clyybber> disruptek yeah but nimph is so big rn |
| 01:42:12 | FromDiscord | <Clyybber> it hurts to even build it |
| 01:42:25 | disruptek | well, it's not even possible to build 2.0. |
| 01:42:36 | disruptek | but, when it was, it was blisteringly fast. |
| 01:42:48 | disruptek | without nimble it's just amazing how quick it is. |
| 01:43:08 | disruptek | for large projects, it went from like 3-4s to 100-200ms. |
| 01:44:02 | FromDiscord | <Varriount> leorize: Yes, but that's a rather verbose solution, and (for safety) must be used virtually everywhere. |
| 01:44:05 | disruptek | it'll be bundled in gitnim anyway. no one will have to build it. |
| 01:44:29 | leorize | @Varriount not really |
| 01:44:55 | * | apahl quit (Ping timeout: 240 seconds) |
| 01:45:34 | leorize | you typically don't abuse command substitution in scripting, and if you do have to use them you'd generally assign them into a variable for further processing |
| 01:46:20 | FromDiscord | <Varriount> leorize: You don't "have" to abuse it, all that needs to happen is for a subcommand to emit odd output. |
| 01:46:40 | FromDiscord | <Varriount> I would also feel better if the tutorial mentioned that idiom. |
| 01:46:57 | leorize | that's what I meant by abusing |
| 01:47:12 | * | apahl joined #nim |
| 01:47:43 | leorize | generally you avoid spawning commands at all costs |
| 01:48:29 | FromDiscord | <Varriount> Generally the purpose of a shell is to spawn commands. |
| 01:48:56 | leorize | if you found yourself frequently using subcommands and passing the output directly to something else as an argument then you might not be doing it right |
| 01:49:12 | disruptek | i agree, it sounds bizarre to me. |
| 01:50:16 | disruptek | i treat the shell like i treat a back alley blowjob. |
| 01:51:56 | FromDiscord | <Varriount> It's still a danger waiting to trip people up, especially since the tutorial/main documentation page doesn't mention it. |
| 01:52:23 | leorize | true that, they're working on better exposing the API |
| 01:52:47 | leorize | they might just stop being stubborn and make `$()` a thing in fish |
| 01:52:56 | disruptek | arg. |
| 01:53:41 | FromDiscord | <Varriount> Then again, I've had to go over bash scripts (not written by me) that were full of unquoted variable and command substitutions, so maybe it's pointless either way. |
| 01:54:02 | leorize | at least you have shellcheck for those :P |
| 01:54:34 | FromDiscord | <Varriount> I treat Bash like an unpredictable power tool. Dangerous, with spinning blades that could possibly fly off at any moment. |
| 01:54:37 | leorize | though shellcheck can't catch errors like calling `sed` and `awk` and expecting them to be the GNU version |
| 01:55:38 | FromDiscord | <Varriount> These days I generally just call perl. It may not be as succinct, but (oddly enough) its generally installed on the systems I work on. |
| 01:55:46 | leorize | when I write shell scripts I assume that the only thing I have are the built-ins and the command I'm scripting around |
| 01:57:04 | disruptek | it sounds like i'm the one having the most fun. |
| 01:57:31 | disruptek | python is at least as portable as perl these days, ime. |
| 01:57:33 | leorize | I found bash parameter expansion works well enough that I can get rid the need for the various text utilities |
| 01:58:01 | disruptek | perl is faster but i'd rather read python. |
| 01:58:52 | FromDiscord | <Varriount> leorize: Do you know if fish has an equivalent to `set -e` and `set -u`? (exit on command exit code > 0, exit on undefined variable reference). |
| 01:59:29 | FromDiscord | <Varriount> Or do I have to suffix everything with `|| exit(1)`? |
| 01:59:56 | disruptek | itym ; or exit 1 |
| 02:00:23 | FromDiscord | <Varriount> Yes |
| 02:00:46 | disruptek | a little beastiality never hurt /anyone/ |
| 02:01:04 | FromDiscord | <Varriount> Huh? |
| 02:06:13 | leorize | @Varriount nope they don't have that |
| 02:06:30 | leorize | I generally just put `if`s on everything |
| 02:07:08 | FromDiscord | <Varriount> Ick |
| 02:07:50 | disruptek | build-compiler ; or exit |
| 02:07:50 | disruptek | and build-sanity ; or exit |
| 02:07:51 | disruptek | and build-test ; or exit |
| 02:08:21 | leorize | you can just use `and` all the way :P |
| 02:08:35 | disruptek | i know, but he wants some codes. |
| 02:08:42 | FromDiscord | <Varriount> https://github.com/fish-shell/fish-shell/issues/510 |
| 02:08:44 | disbot | ➥ Option to abort a script if a command fails (e.g., set -o errexit) ; snippet at 12https://play.nim-lang.org/#ix=2uXO |
| 02:09:37 | FromDiscord | <Varriount> I should just implement my own unholy cross between ansible and fish |
| 02:09:58 | leorize | make nim shell happen |
| 02:10:22 | FromDiscord | <Varriount> Well, I already have a base: https://github.com/Varriount/commandant |
| 02:10:26 | leorize | there's xonsh now for python-like stuff but the ergonomic is not nice |
| 02:11:34 | leorize | https://skarnet.org/software/execline/ <- you might be interested in that one |
| 02:11:44 | leorize | it's the truest sense of everything is a command |
| 02:11:49 | FromDiscord | <Varriount> Although, how Ansible became popular is beyond me. Any advanced variable handling requires mental gymnastics |
| 02:12:42 | FromDiscord | <Varriount> ooh, thanks. That looks quite interesting. |
| 02:13:50 | disruptek | i guess you looked at oil shell. |
| 02:14:05 | disruptek | the one i was interested in was more lispy but i can't remember what it's called. |
| 02:14:08 | FromDiscord | <Varriount> I've seen it. I've also looked at Elvish (Go shell) |
| 02:14:24 | FromDiscord | <Varriount> That might be elvish |
| 02:14:38 | disruptek | i think so. |
| 02:15:41 | leorize | 99% of the time you only need a good enough shell for the job |
| 02:15:51 | leorize | anything more complicated you write them in Nim |
| 02:15:57 | disruptek | yeah, the thing is that portability is more important. |
| 02:16:07 | leorize | now the only issue is that osproc is terrible |
| 02:16:09 | disruptek | i know i can get fish where i need it, and i'd rather write nim. |
| 02:16:26 | disruptek | i don't find it to be a problem, honestly. |
| 02:16:32 | disruptek | it's annoying sometimes, but whatever. |
| 02:17:07 | leorize | some of my use cases require seamless piping, redirecting and controlled fd passing |
| 02:17:11 | disruptek | to put that a different way, when i look at my osproc code, it's ugly. on the other hand, i don't have to look at it often. |
| 02:17:13 | leorize | all of which osproc can't do |
| 02:17:30 | leorize | I'm brewing my own replacement though |
| 02:17:38 | disruptek | yeah, if i'm going out to osproc i'm already lubed up. |
| 02:17:51 | leorize | hopefully I can get it to good enough so I can toss it out there |
| 02:18:01 | disruptek | i wouldn't wait. |
| 02:18:42 | FromDiscord | <Varriount> leorize: I'd eventually like to port D's subprocess module to Nim, possibly adding in things like piping |
| 02:18:56 | leorize | I think D's subprocess is inelegant |
| 02:19:24 | leorize | Go's approach is much leaner and simpler |
| 02:19:51 | leorize | Rust appears to take some pages from Go's approach |
| 02:21:07 | disruptek | do you write a lot of go? |
| 02:21:30 | FromDiscord | <Varriount> Hm, Go appears to take the standard "stuff 5 million arguments into a single function call" and translate it into an object-oriented (or rather, struct-oriented) approach. |
| 02:22:14 | FromDiscord | <Varriount> Which, I guess they would have to do, given Go's lack of named parameters (using closure options is possible, but quite inefficient) |
| 02:22:41 | FromDiscord | <Varriount> leorize: Yeah, actually, that looks better. |
| 02:24:12 | FromDiscord | <Varriount> leorize: How do you imagine something like piping output from one process to another would be achieved? Do you think we could take advantage of Nim's operator overloading? |
| 02:27:37 | * | muffindrake quit (Ping timeout: 272 seconds) |
| 02:28:38 | leorize[m] | "disruptek" (https://matrix.to/#/@freenode_disruptek:matrix.org): I don't even write Go :p |
| 02:29:25 | * | muffindrake joined #nim |
| 02:30:56 | * | aenesidemus_ quit (Read error: Connection reset by peer) |
| 02:31:02 | FromDiscord | <Hearthstone> Hmmm... From what I've seen about Nimph |
| 02:31:26 | FromDiscord | <Hearthstone> The 2.0 version will be better then nimble, judging from what I've seen |
| 02:31:29 | FromDiscord | <Hearthstone> Correct? |
| 02:31:45 | * | audiofile joined #nim |
| 02:31:55 | disruptek | i mean, i think the 1.0 version is better, but i'm extremely biased. |
| 02:32:11 | FromDiscord | <Hearthstone> Because you made it? |
| 02:32:27 | disruptek | yes, and i made it to my spec. |
| 02:32:52 | FromDiscord | <Hearthstone> Hm |
| 02:33:07 | FromDiscord | <Hearthstone> How does importing modules work in Nimph? |
| 02:33:09 | disruptek | custom software is pretty nice to use. |
| 02:33:24 | FromDiscord | <Hearthstone> And can it install everything Nimble can install? |
| 02:33:34 | disruptek | sure. |
| 02:33:59 | FromDiscord | <Hearthstone> Hm |
| 02:34:15 | disruptek | i can't imagine not writing my own software. |
| 02:34:32 | FromDiscord | <Hearthstone> Would importing modules on nimble work like golang's method? Like `github.com/user/repo`? |
| 02:34:32 | disruptek | i guess i /do/ know what i wanna be when i grow up! |
| 02:34:34 | FromDiscord | <Hearthstone> Lmao |
| 02:34:59 | disruptek | there is a milestone for pragma-based package management. |
| 02:35:22 | disruptek | i haven't decided if it'll go in before nim-2.0 but it'll definitely go in by then. |
| 02:35:24 | FromDiscord | <Hearthstone> Oh? |
| 02:36:11 | disruptek | {.nimph: "https://github.com/disruptek/testes#2.0.3".} will be a thing, yes. |
| 02:36:17 | FromDiscord | <Hearthstone> Oo |
| 02:36:37 | FromDiscord | <Hearthstone> And then the library will be imported in that proc? |
| 02:36:57 | disruptek | i dunno the specifics; it's not really important. |
| 02:37:03 | disruptek | magic will happen. |
| 02:37:11 | disruptek | then you'll be able to play with my testes... |
| 02:37:25 | disruptek | some say that's when the magic /really/ happens. |
| 02:37:29 | FromDiscord | <Hearthstone> Lmao |
| 02:37:35 | FromDiscord | <Hearthstone> Hmm |
| 02:37:58 | disruptek | the direction we're moving in is to put everything into the source. |
| 02:38:04 | FromDiscord | <Hearthstone> If something like `import "github.com/user/repo"` becomes a thing let me know- |
| 02:38:22 | FromDiscord | <Hearthstone> Because that'll be fun xD |
| 02:38:47 | disruptek | it's not really that different from what we have now. |
| 02:39:07 | leorize[m] | @Varriount I think Go just nail this well since it got people who have strong bg in the Unix philosophy as core devs. |
| 02:39:07 | disruptek | requires "https://github.com/disruptek/testes#2.0.3" |
| 02:39:25 | leorize[m] | about operators, I'm not sure tbh |
| 02:39:48 | FromDiscord | <Hearthstone> Oh? |
| 02:39:52 | FromDiscord | <Hearthstone> Is that a thing? |
| 02:39:58 | disruptek | yes. |
| 02:40:11 | FromDiscord | <Hearthstone> Does that download the required lib on the go? |
| 02:40:18 | disruptek | no. |
| 02:40:26 | FromDiscord | <Hearthstone> How does it work then? |
| 02:40:38 | disruptek | that's syntax for your .nimble file. nimble or nimph will use it to install the package. |
| 02:41:30 | FromDiscord | <Hearthstone> Oh alright |
| 02:41:40 | FromDiscord | <Hearthstone> I mean in the .nim files |
| 02:41:48 | disruptek | nah, that's not a thing there. |
| 02:41:56 | FromDiscord | <Hearthstone> I wanna do something similar to what go does for most imports because why not- |
| 02:42:03 | FromDiscord | <Hearthstone> Alright then :P |
| 02:42:20 | disruptek | it's a little verbose. that's a reason not to. |
| 02:42:43 | FromDiscord | <Hearthstone> Really? |
| 02:42:51 | disruptek | i think araq is pressured not to make it hard for airgapped folks to use nim, so that's another reason he resists it. |
| 02:43:07 | disruptek | `import testes` is shorter than any of these URLs. |
| 02:43:12 | FromDiscord | <Hearthstone> Wdym by airgapped? |
| 02:43:17 | FromDiscord | <Hearthstone> Definitely, that's true |
| 02:43:25 | disruptek | networks which are not on the internet. |
| 02:43:44 | FromDiscord | <Hearthstone> Oh alright |
| 02:44:05 | disruptek | think fintech or gov. |
| 02:44:58 | FromDiscord | <Hearthstone> Makes sense |
| 02:47:49 | FromDiscord | <Varriount> leorize: I write Go. The standard library is... ok. It's a bit anemic in some unexpected places |
| 02:48:25 | * | thomasross quit (Ping timeout: 240 seconds) |
| 02:48:25 | FromDiscord | <Varriount> Though, possibly that's due to the language, which I find a bit too conservative. |
| 02:48:48 | shashlick | disruptek: what's new with nimph |
| 02:48:57 | FromDiscord | <Varriount> Something like Nim's templates would "go" a long way to fixing my gripes with it. |
| 02:49:14 | disruptek | i'm doing a second version. |
| 02:49:41 | disruptek | currently trying to set fire to my desk with concept recursion. |
| 02:50:32 | shashlick | reading your pr |
| 02:52:07 | disruptek | yeah, it doesn't build yet. |
| 02:52:33 | shashlick | not the code, the description |
| 02:52:47 | shashlick | so I might steal your nim.cfg writing code if you are going to do that |
| 02:52:48 | disruptek | ah, yeah. |
| 02:53:07 | disruptek | that code already exists in nimph/config.nim |
| 02:53:09 | shashlick | I was going to work on that next but I'm lazier than you |
| 02:53:19 | shashlick | what all do you write to nim.cfg |
| 02:53:30 | shashlick | and do you work with existing nim.cfg |
| 02:53:37 | shashlick | user modified |
| 02:53:41 | disruptek | you will have to rip it apart though because i write a new file, then parse it with the compiler to make sure it's kosher, then move it into place. |
| 02:53:49 | disruptek | yeah, it modifies the existing file. |
| 02:54:03 | shashlick | sounds about what I was thinking of doing |
| 02:54:21 | shashlick | do you put stuff in a separate "modified by nimph" section |
| 02:54:24 | shashlick | or just inline |
| 02:54:27 | disruptek | one of the very few bugs i had was that people on windows created nim.cfg files that had no trailing newline. |
| 02:54:39 | disruptek | i was adding lines. this corrupted the file. |
| 02:54:50 | disruptek | so i put a comment line in to solve that problem. |
| 02:55:01 | disruptek | just a heads-up on something you will want to think about. |
| 02:55:06 | shashlick | cool |
| 02:55:16 | shashlick | I have some concerns with writing nim.cfg though |
| 02:55:19 | disruptek | other than that it doesn't do anything special. |
| 02:55:24 | shashlick | since you have done it already, want to run it by you |
| 02:55:35 | disruptek | i don't compose it from scratch. |
| 02:55:37 | disruptek | i just append. |
| 02:55:41 | shashlick | if you are putting --path or --nimblePath into it |
| 02:55:42 | disruptek | oh, and i remove stuff. |
| 02:55:57 | shashlick | and we are in localdeps mode |
| 02:56:00 | disruptek | well, nimph 2 will just compose it from scratch. |
| 02:56:16 | shashlick | the paths will be $project/nimbledeps/pkgs/xyz |
| 02:56:38 | shashlick | which means nim.cfg cannot be checked in since localdeps isn't really something you check in |
| 02:56:48 | shashlick | in this case the paths |
| 02:56:58 | disruptek | nim.cfg never gets checked-in, but you can check in the project.nim.cfg. |
| 02:57:19 | shashlick | I guess the consolation is that these paths won't exist on users systems so Nim will just ignore them and they will be cruft if checked in |
| 02:57:29 | shashlick | some users do checkin their nim.cfg |
| 02:57:41 | shashlick | and if we are adding to it, we will have to consider that no? |
| 02:57:42 | disruptek | those people will be first against the wall. |
| 02:58:01 | disruptek | no; we decided the pm will blow it away, remember? |
| 02:58:05 | shashlick | problem with project.nim.cfg is that it only affects project.nim |
| 02:58:17 | shashlick | some people use nim.cfg to affect their test cases and multiple binaries |
| 02:58:21 | disruptek | i'm okay with that. |
| 02:58:52 | disruptek | they can use config.nims, too, if they must. |
| 02:59:14 | leorize | @Varriount I guess you can expect that any constructs imported from Go to Nim will be a few dozen times more flexible :P |
| 02:59:15 | disruptek | it's a file that is even harder for us to parse. or easier, depending on how you look at it. |
| 02:59:17 | shashlick | it is an education issue then - as it is I'm hearing crickets with my nimble changes it concerns me |
| 02:59:40 | disruptek | remember that there aren't that many nim users. |
| 02:59:53 | disruptek | there are like 11 of us. 13 if you include zevv. |
| 03:00:22 | shashlick | well, the fact that 4raq only wants nimble updates to go out with Nim releases, broken nimble means a new nim release |
| 03:00:39 | shashlick | he gets counted twice for his two v's? |
| 03:00:53 | shashlick | I have two h's |
| 03:00:53 | disruptek | each hand is like a full nimion. |
| 03:01:16 | disruptek | when did nimble become broken? |
| 03:01:40 | shashlick | well, if my so called fixes/enhancements break the community |
| 03:01:43 | disruptek | it worked for me with your pr. |
| 03:02:08 | disruptek | you won't know /until/ the next nim release. |
| 03:02:29 | shashlick | that's my point - I'd rather have nimble release as and when, independent of nim |
| 03:02:41 | disruptek | i agree. |
| 03:02:42 | shashlick | already nimble is tested with Nim as far back as 0.20.2 |
| 03:02:49 | disruptek | maybe choosenim can bump it. |
| 03:02:52 | shashlick | but 4raq is not a fan of that |
| 03:03:16 | shashlick | you can already do `koch nimble --latest` or `nimble install nimble` |
| 03:03:16 | disruptek | well, i agree with him. 1.0+ is the only thing that makes sense. |
| 03:03:27 | disruptek | but no one will do that. |
| 03:03:34 | shashlick | exactly |
| 03:03:35 | disruptek | choosenim will get a newer nimble out faster. |
| 03:03:49 | disruptek | then you can continue to ship an older nimble with nim as long as you want. |
| 03:04:05 | shashlick | why choosenim instead of nimble |
| 03:04:13 | disruptek | because no one will update nimble. |
| 03:04:19 | disruptek | but they will use choosenim to update nim. |
| 03:04:35 | * | PMunch quit (Quit: leaving) |
| 03:05:23 | shashlick | hmm |
| 03:05:46 | leorize | shashlick: are you using CI to test nimble? |
| 03:07:27 | shashlick | ya CI is all green |
| 03:09:55 | shashlick | just that new features aren't getting much usage thru the devel cycle |
| 03:10:02 | shashlick | to expose bugs before 1.4 comes out |
| 03:10:43 | disruptek | i don't think it's a problem. |
| 03:11:02 | disruptek | 1.4.2 can come out whenever. |
| 03:39:13 | FromDiscord | <Elegant Beef> Is `0..^10` still a valid iterator? |
| 03:39:20 | disruptek | sure. |
| 03:39:54 | FromDiscord | <Elegant Beef> Well it doesnt work |
| 03:40:06 | disruptek | what do you expect? |
| 03:40:13 | FromDiscord | <Elegant Beef> it to work based off the tutorial |
| 03:40:31 | disruptek | paste the line of code in which you use it. |
| 03:40:45 | FromDiscord | <Elegant Beef> !eval for x in 0..^10: echo x |
| 03:40:46 | NimBot | Compile failed: /usercode/in.nim(1, 11) Error: type mismatch: got <HSlice[system.int, system.BackwardsIndex]> |
| 03:41:16 | disruptek | yeah, it's used as an index into an array/seq. |
| 03:41:16 | Prestige | I think you wan 0..<10 |
| 03:41:22 | Prestige | want* |
| 03:41:23 | disruptek | most likely, yes. |
| 03:43:00 | * | bung joined #nim |
| 03:43:11 | FromDiscord | <Elegant Beef> I mean the tutorial indicates it's a thing so i was curious if it's supposed to work https://media.discordapp.net/attachments/371759389889003532/746937555282427985/unknown.png |
| 03:43:41 | disruptek | well, it's not an index, is it? |
| 03:44:04 | disruptek | i think that's poorly-worded at best. |
| 03:44:22 | disruptek | ^ is the backwards index operator. |
| 03:44:41 | disruptek | you apply it to an index to indicate that you want the index N-less from the end. |
| 03:44:51 | disruptek | x[^1] is the last index. |
| 03:44:56 | FromDiscord | <Elegant Beef> Yea i know that, i was just curious if it was supposed to work |
| 03:44:58 | disruptek | x[^2] is the second-to-last index. |
| 03:45:27 | disruptek | !eval for i in 0 ..^ 10: echo i |
| 03:45:28 | NimBot | Compile failed: /usercode/in.nim(1, 12) Error: type mismatch: got <HSlice[system.int, system.BackwardsIndex]> |
| 03:46:05 | disruptek | i dunno, you're making me wonder if it was removed recently and i just never learned it. |
| 03:46:30 | disruptek | omfg the hax0r is back. |
| 03:47:12 | disruptek | !eval for i in 0 ..< 10: echo i |
| 03:47:14 | NimBot | 0↵1↵2↵3↵4↵5↵6↵7↵8↵9 |
| 03:47:50 | disruptek | i have a feeling it existed and was removed. |
| 03:48:14 | FromDiscord | <Elegant Beef> That was my intuition aswell |
| 03:48:22 | disruptek | the problem with my concepts is that i apparently cannot use a concept to define a concept. |
| 03:48:32 | disruptek | i mean, the same concept cannot be used to define itself. |
| 03:48:46 | disruptek | ain't that some booolsheeit? |
| 03:48:55 | disruptek | so to speak. |
| 04:02:53 | * | vicfred quit (Quit: Leaving) |
| 04:05:12 | * | vicfred joined #nim |
| 04:06:01 | * | bung quit (Ping timeout: 258 seconds) |
| 04:06:01 | * | supakeen quit (Quit: WeeChat 2.9) |
| 04:06:35 | * | supakeen joined #nim |
| 04:13:13 | * | njoseph quit (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) |
| 04:13:21 | * | njoseph joined #nim |
| 04:18:30 | * | bung joined #nim |
| 04:40:31 | * | solitudesf joined #nim |
| 04:43:25 | * | audiofile quit (Ping timeout: 240 seconds) |
| 05:03:11 | * | NimBot joined #nim |
| 05:14:45 | * | letto quit (Quit: Konversation terminated!) |
| 05:15:29 | * | letto joined #nim |
| 05:50:11 | * | arecacea1 quit (Remote host closed the connection) |
| 05:50:40 | * | arecacea1 joined #nim |
| 05:51:53 | * | Zectbumo joined #nim |
| 06:37:25 | * | Cthalupa quit (Ping timeout: 240 seconds) |
| 06:38:28 | * | Cthalupa joined #nim |
| 06:56:35 | * | icyphox joined #nim |
| 06:56:55 | icyphox | What's a good way to parse multipart/form-data in Nim? |
| 07:00:33 | FromDiscord | <vieru> hi |
| 07:00:50 | FromDiscord | <vieru> i still cant get openssl with static linking to work :/ |
| 07:03:38 | FromDiscord | <Rika> maybe try https://github.com/kaushalmodi/hello_musl ? |
| 07:05:29 | FromDiscord | <vieru> isn't this linux only ? |
| 07:08:16 | FromDiscord | <Rika> musl can be used on windows |
| 07:12:57 | FromDiscord | <vieru> do i just download it or how can i install it ? |
| 07:13:10 | FromDiscord | <vieru> sry for those noob questions |
| 07:16:19 | FromDiscord | <Rika> i cant help there since i dont use windows anymore |
| 07:16:38 | FromDiscord | <Rika> (i dont know) |
| 07:17:41 | FromDiscord | <vieru> oh well fuck |
| 07:17:50 | FromDiscord | <vieru> ill just click things till they work |
| 07:32:58 | eery | musl is linux only. It would probably work within WSL(2), though. Not with win32/msvc |
| 07:33:18 | * | k0mpjut0r joined #nim |
| 07:38:38 | FromDiscord | <vieru> now im trying https://github.com/genotrance/nimssl↵mby this will help |
| 07:57:33 | * | hnOsmium0001 quit (Quit: Connection closed for inactivity) |
| 08:17:08 | * | fowl_ quit (Read error: Connection reset by peer) |
| 08:17:12 | * | fowl quit (Read error: Connection reset by peer) |
| 08:17:39 | * | kwilczynski quit (Read error: Connection reset by peer) |
| 08:17:57 | * | sirn quit (Read error: Connection reset by peer) |
| 08:20:09 | * | sirn joined #nim |
| 08:21:21 | * | kwilczynski joined #nim |
| 08:22:15 | * | fowl_ joined #nim |
| 08:22:20 | * | fowl joined #nim |
| 08:59:40 | FromGitter | <gogolxdong> Anyone works on developping Android App in Nim? |
| 09:00:15 | FromGitter | <gogolxdong> Developping on mobile phone in general. |
| 09:00:23 | * | reversem3 quit (Quit: Idle for 30+ days) |
| 09:02:11 | FromGitter | <gogolxdong> Noticed dali, hellomello, nim-adb, want to run a http server on mobile phone as simply as possible, have to build app with React Native or Flutter for now. |
| 09:02:47 | FromGitter | <gogolxdong> They are too heavy. |
| 09:03:27 | * | drewr quit (Ping timeout: 244 seconds) |
| 09:04:10 | Zevv | consider running in termux, tis pretty lightweight and very unixy |
| 09:04:32 | Zevv | but its not really "apps" your making then |
| 09:05:07 | * | rridley[m] quit (Quit: Idle for 30+ days) |
| 09:05:07 | * | swamptest1[m] quit (Quit: Idle for 30+ days) |
| 09:05:56 | * | drewr joined #nim |
| 09:09:30 | FromDiscord | <jseb> hello, is is possible to use Jester (with nginx as proxy) to upload very big files ? (say, 1 gigabyte)↵I have read the forum and it seems the correct method would be to use Jester with websockets |
| 09:09:46 | FromGitter | <gogolxdong> Heard that the linux kernel Android opensource project doesn't use glibc, so there is much trouble to run c-like thing on Android. |
| 09:10:49 | FromGitter | <gogolxdong> the linux kernel Android based on. |
| 09:11:02 | FromDiscord | <Yardanico> No, it's not about the kernel |
| 09:11:15 | FromDiscord | <Yardanico> Kernel has nothing to do with the C library |
| 09:11:19 | FromGitter | <gogolxdong> Fee like want to write a mobile phone OS |
| 09:11:28 | FromDiscord | <Yardanico> But yes, Android uses bionic |
| 09:11:56 | FromDiscord | <Yardanico> But no one stops you from statically linking with musl for some terminal applications |
| 09:13:39 | FromGitter | <gogolxdong> It sounds tricky to me. Can you run a http server directly on mobile phone, or have to work within Android context? |
| 09:16:34 | FromGitter | <gogolxdong> Has Ubuntu mobile OS failed? |
| 09:17:46 | * | Zectbumo quit (Remote host closed the connection) |
| 09:18:15 | FromDiscord | <Elegant Beef> Ubuntu Ports exist |
| 09:18:16 | FromDiscord | <Elegant Beef> (edit) 'exist' => 'exists' |
| 09:21:34 | FromGitter | <gogolxdong> Is it still working? The most important thing is the baseband driver for 5G chipsets. They are integrated with mobile OS. |
| 09:23:13 | FromDiscord | <Elegant Beef> It only supports specific phones offically |
| 09:24:31 | FromGitter | <gogolxdong> It reads as only supports specific 5G baseband chipset. |
| 09:31:54 | * | icyphox quit (Quit: catch you on the flipside :^)) |
| 09:40:44 | * | icyphox joined #nim |
| 10:07:37 | Zevv | gogolxdong: bionic is a fairly standard libc, so thats not your problem |
| 10:08:09 | Zevv | you can write normal c or c++ code and run that as android service, no problem |
| 10:08:38 | Zevv | but you cant make c++-only apps though, there alway needs to be the bare java stuff to make it an "app" in the android sense |
| 10:19:47 | * | krux02 joined #nim |
| 10:24:42 | FromDiscord | <kodkuce> hmm, c# like static classes in nim would be like just make some object and set it as global variable right? |
| 10:25:36 | FromGitter | <gogolxdong> no need to make app considering what I need is only a http server running on ipv6 address of mobile phone, how can I achieve that? |
| 10:26:01 | Zevv | like i said: termux is your easiest way to go |
| 10:26:23 | Zevv | if you want to make an android service type server, you need root to be able to install & run it |
| 10:26:29 | Zevv | for termux, you can do it all as a user |
| 10:26:36 | Zevv | and have a friendly development environment |
| 10:32:02 | FromGitter | <gogolxdong> Is it Android NDK? |
| 10:32:15 | FromGitter | <gogolxdong> The Android NDK is a toolset that lets you implement parts of your app in native code, using languages such as C and C++. For certain types of apps, this can help you reuse code libraries written in those languages. |
| 10:32:31 | FromGitter | <gogolxdong> I think it's an android service type server. |
| 10:32:53 | Zevv | termux is not ndk |
| 10:33:17 | Zevv | you can do ndk, but like i said, you still need some java |
| 10:47:47 | * | ja0nz joined #nim |
| 10:57:22 | * | bouzu_ joined #nim |
| 11:09:21 | FromDiscord | <lakmatiol> sent a code paste, see https://play.nim-lang.org/#ix=2uZK |
| 11:12:59 | FromDiscord | <lqdev> sent a code paste, see https://play.nim-lang.org/#ix=2uZM |
| 11:13:14 | * | apahl quit (Ping timeout: 244 seconds) |
| 11:13:24 | FromDiscord | <lakmatiol> ah, that is quite nice |
| 11:13:46 | FromDiscord | <lqdev> you can always create a custom iterator out of it if you want to |
| 11:14:25 | * | apahl joined #nim |
| 11:18:22 | FromDiscord | <lakmatiol> how do I create a range backwards, like `10..0` |
| 11:19:26 | FromDiscord | <lakmatiol> ah, countdown |
| 11:21:03 | * | tane joined #nim |
| 11:30:26 | * | lritter joined #nim |
| 11:36:39 | bung | how to check Duration is negtive ? |
| 11:37:43 | FromDiscord | <Rika> if it is less than `DurationZero` |
| 11:37:54 | FromDiscord | <Rika> (DurationZero can be found in times module i think) |
| 11:38:15 | FromDiscord | <Rika> yeah https://nim-lang.org/docs/times.html#DurationZero |
| 11:38:23 | bung | yeah there is , thank you! |
| 11:40:20 | * | fredrikhr quit (Read error: Connection reset by peer) |
| 11:40:53 | * | fredrikhr joined #nim |
| 11:46:03 | * | fredrikhr quit (Read error: Connection reset by peer) |
| 11:46:23 | * | fredrikhr joined #nim |
| 11:56:51 | FromDiscord | <vieru> i still can't get openssl with static linking to work :/ |
| 11:59:46 | leorize | you're on windows, correct? |
| 11:59:52 | FromDiscord | <vieru> yep |
| 11:59:52 | leorize | how are you doing the static linking? |
| 12:00:06 | FromDiscord | <vieru> `nim c -d:ssl -p:. --dynlibOverride:ssl --passL:-Lsrc\\lib\\more\\openssl\\lib --passL:-lcrypto --passL:-lssl -r tssl.nim` |
| 12:00:10 | FromDiscord | <vieru> like this |
| 12:00:19 | FromDiscord | <vieru> is this what youre referring to ? |
| 12:00:24 | leorize | yep |
| 12:00:38 | leorize | try forward slash instead |
| 12:00:47 | FromDiscord | <vieru> 1sec |
| 12:01:34 | FromDiscord | <vieru> doesn't make any difference |
| 12:01:37 | FromDiscord | <vieru> ```↵could not import: OpenSSL_version_num↵Error: execution of an external program failed↵``` |
| 12:02:10 | leorize | what version of openssl do you have? |
| 12:02:30 | FromDiscord | <vieru> i think its 1.1.1g |
| 12:02:36 | FromDiscord | <vieru> lemme check |
| 12:03:06 | * | fredrikhr quit (Read error: Connection reset by peer) |
| 12:03:20 | * | fredrikhr joined #nim |
| 12:03:28 | FromDiscord | <vieru> openssl-1.1.1g-win64-mingw |
| 12:03:45 | FromDiscord | <vieru> i've downloaded a precompiled one |
| 12:03:54 | FromDiscord | <vieru> cus i can't get perl to work |
| 12:03:59 | FromDiscord | <vieru> to compile my own version |
| 12:04:17 | leorize | add --dynlibOverride:crypto |
| 12:04:53 | FromDiscord | <vieru> like this ? |
| 12:04:56 | FromDiscord | <vieru> ```↵nim c -d:ssl -p:. --dynlibOverride:ssl --dynlibOverride:crypto --passL:-Lsrc/lib/more/openssl/lib --passL:-lcrypto --passL:-lssl -r tssl.nim↵``` |
| 12:05:00 | leorize | yep |
| 12:05:07 | FromDiscord | <vieru> same output |
| 12:05:33 | leorize | try --dynlibOverrideAll -d:noOpenSSLHacks then |
| 12:06:02 | * | supakeen quit (Quit: WeeChat 2.9) |
| 12:06:43 | * | supakeen joined #nim |
| 12:06:44 | FromDiscord | <vieru> should i keep `--dynlibOverride:ssl --dynlibOverride:crypto` ? |
| 12:07:03 | leorize | you don't have to keep those |
| 12:07:09 | leorize | not that it hurts to have them :P |
| 12:07:51 | FromDiscord | <vieru> i've got some big error thingy |
| 12:07:57 | FromDiscord | <vieru> imma do -f |
| 12:08:02 | FromDiscord | <vieru> mby it fixes it |
| 12:08:10 | leorize | that sounds like progress :P |
| 12:08:37 | FromDiscord | <vieru> https://hastebin.com/eriyiqeqig.cs |
| 12:08:44 | FromDiscord | <vieru> that's the error |
| 12:08:57 | * | luis_ joined #nim |
| 12:09:03 | leorize | oh yea ofc that won't work, you're on windows... |
| 12:09:21 | * | PMunch joined #nim |
| 12:09:24 | leorize | let's try this instead, keep the dynliboverrides, then remove all `passL` |
| 12:09:49 | leorize | if it couldn't link due to missing symbols then we know that the overrides are working |
| 12:10:19 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. --dynlibOverride:ssl --dynlibOverride:crypto -r tssl.nim` |
| 12:10:34 | leorize | yea |
| 12:10:58 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v00 |
| 12:11:14 | FromDiscord | <vieru> same output every fucking time |
| 12:11:36 | leorize | looks like the overrides are being ignored |
| 12:12:41 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v01 |
| 12:12:42 | FromDiscord | <vieru> if it helps |
| 12:12:56 | FromDiscord | <vieru> gcc if from nim's website |
| 12:13:00 | FromDiscord | <vieru> (edit) 'if' => 'is' |
| 12:13:25 | luis_ | Hi all, I am trying to convert c code into nim with c2nim, the tool complains of ; missing: /home/luis/Documents/OneDrive/Coding/Nim/Learning/miniToxNim/minitoxnim.c(68, 55) Error: expected ' |
| 12:13:27 | luis_ | ;' |
| 12:13:50 | luis_ | sorry, the error is Error: expected ';' |
| 12:13:58 | luis_ | how can I debug this? |
| 12:14:16 | luis_ | seems to be simple, but I have no idea on C api |
| 12:14:45 | luis_ | and this is the file: https://github.com/hqwrong/minitox/blob/master/minitox.c |
| 12:16:11 | FromGitter | <kingramx> Hello Guys |
| 12:16:11 | * | fredrikhr quit (Read error: Connection reset by peer) |
| 12:16:39 | leorize | @vieru looks like the dynlibOverride algo isn't tuned for windows... let me see if I can produce an override string that'd trigger it |
| 12:16:55 | * | fredrikhr joined #nim |
| 12:17:39 | FromDiscord | <vieru> thank you so much↵this bs had me pulling my hair |
| 12:21:35 | * | fredrikhr quit (Read error: Connection reset by peer) |
| 12:21:49 | * | fredrikhr joined #nim |
| 12:23:01 | * | luis_ quit (Quit: luis_) |
| 12:23:32 | * | luis_ joined #nim |
| 12:23:49 | leorize | ok this is a terrible way but maybe it'd work: --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" |
| 12:24:02 | leorize | mind the trailing slashes |
| 12:24:37 | FromDiscord | <vieru> ok imma try it rn |
| 12:24:48 | FromDiscord | <ache of head> luis_: this doesn’t seem to be the file the error is referencing. |
| 12:25:08 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. --dynlibOverride:ssl --dynlibOverride:crypto -d:sslVersion:"(" -r tssl.nim` |
| 12:25:32 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v06 |
| 12:25:44 | leorize | the `-` at the end of `ssl` and `crypto` matters, please add them |
| 12:26:06 | FromDiscord | <vieru> oh f |
| 12:26:51 | luis_ | I renamed it a bit... |
| 12:26:54 | FromDiscord | <vieru> https://hastebin.com/gizakotere.coffeescript |
| 12:27:19 | luis_ | ache of head: I renamed it a bit |
| 12:27:53 | leorize | luis_: see #def and #pp: https://github.com/nim-lang/c2nim/blob/master/doc/c2nim.rst#def-directive |
| 12:28:07 | leorize | c2nim don't evaluate C macros by default |
| 12:28:25 | luis_ | oh, ok, thanks |
| 12:28:31 | leorize | @vieru: it's working, now add the passL back |
| 12:29:16 | leorize | you might want to do it like this --passL:-L/path/to/ssl --passL:-Bstatic --passL:-lssl --passL:-lcrypto --passL:-Bdynamic |
| 12:29:17 | * | rockcavera quit (Remote host closed the connection) |
| 12:29:21 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v09 |
| 12:29:37 | FromDiscord | <vieru> i ran↵`nim c -f -d:ssl -p:. --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:-Lsrc/lib/more/openssl/lib --passL:-lcrypto --passL:-lssl -r tssl.nim` |
| 12:30:28 | FromDiscord | <vieru> after i ran ↵`nim c -f -d:ssl -p:. --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:-Lsrc/lib/more/openssl/lib --passL:-Bstatic --passL:-lssl --passL:-lcrypto --passL:-Bdynamic -r tssl.nim` |
| 12:30:36 | FromDiscord | <vieru> i got this |
| 12:30:41 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v09 |
| 12:30:52 | leorize | add -d:noOpenSSLHacks |
| 12:31:03 | * | waleee-cl joined #nim |
| 12:31:15 | FromDiscord | <vieru> before dynlib overrides is ok ? |
| 12:31:22 | leorize | yea |
| 12:31:22 | FromDiscord | <vieru> or it doesnt matter |
| 12:31:32 | leorize | it doesn't matter |
| 12:32:00 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:-Lsrc/lib/more/openssl/lib --passL:-Bstatic --passL:-lssl --passL:-lcrypto --passL:-Bdynamic -r tssl.nim` |
| 12:32:11 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v0b |
| 12:32:12 | FromDiscord | <vieru> this worked |
| 12:32:29 | leorize | please file a bug report on how absurd it is to get this working on windows :P |
| 12:32:57 | FromDiscord | <vieru> on github ? |
| 12:33:02 | leorize | yep |
| 12:33:19 | FromDiscord | <vieru> ok imma try |
| 12:33:23 | FromDiscord | <vieru> thank you |
| 12:33:46 | leorize | np :) |
| 12:34:12 | * | evilkhaoskat joined #nim |
| 12:34:17 | FromDiscord | <vieru> what's you gh↵i want to tag you |
| 12:34:25 | FromDiscord | <vieru> you found this |
| 12:34:27 | leorize | alaviss is the name |
| 12:34:58 | FromDiscord | <vieru> thanks again, imma file the report rn |
| 12:35:44 | leorize | people don't usually do static linking on windows, which is probably why it's broken there for so long :P |
| 12:37:51 | FromDiscord | <vieru> that may be the cause↵ive also tried opencv but i couldn't get it to work |
| 12:38:10 | * | evilkhaoskat quit (Client Quit) |
| 12:40:36 | * | euantor is now known as euantorano |
| 12:42:01 | * | euantorano quit () |
| 12:42:15 | * | euantorano joined #nim |
| 12:46:49 | FromDiscord | <vieru> leorize: here we go again |
| 12:46:57 | FromDiscord | <vieru> https://media.discordapp.net/attachments/371759389889003532/747074396551905370/unknown.png |
| 12:47:10 | FromDiscord | <vieru> i've sent tssl.exe to my friend |
| 12:47:27 | leorize | seems like the static linking didn't statically link :P |
| 12:47:33 | FromDiscord | <vieru> yep |
| 12:47:42 | FromDiscord | <vieru> 400k executable size |
| 12:47:45 | FromDiscord | <vieru> (edit) '400k' => '400kb' |
| 12:47:56 | leorize | change your passL to this: `--passL:/path/to/libssl.a --passL:/path/to/libcrypto.a` |
| 12:48:06 | leorize | this should force gcc to do static linking |
| 12:48:18 | FromDiscord | <vieru> ok 1 sec |
| 12:49:48 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:-Lsrc/lib/more/openssl/lib --passL:-Bstatic --passL:src/lib/more/openssl/lib/libssl.a --passL:src/lib/more/openssl/lib/libcrypto.a --passL:-Bdynamic -r tssl.nim` |
| 12:49:53 | FromDiscord | <vieru> is this ok ? |
| 12:50:05 | leorize | yea |
| 12:50:14 | leorize | you no longer need the -L and -Bstatic/dynamic |
| 12:51:25 | FromDiscord | <vieru> i've got this error |
| 12:51:26 | FromDiscord | <vieru> https://hastebin.com/qelixameli.coffeescript |
| 12:51:35 | FromDiscord | <vieru> ill try without -L and -Bstatic/dynamic |
| 12:52:56 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:src/lib/more/openssl/lib/libssl.a --passL:src/lib/more/openssl/lib/libcrypto.a -r tssl.nim`↵↵got me ↵https://hastebin.com/rujanuvoxa.coffeescript |
| 12:54:56 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl- --dynlibOverride:crypto- -d:sslVersion:"(" --passL:src/lib/more/openssl/lib/libssl.dll.a --passL:src/lib/more/openssl/lib/libcrypto.dll.a -r tssl.nim` |
| 12:55:03 | FromDiscord | <vieru> this got me no error |
| 12:55:28 | FromDiscord | <vieru> https://media.discordapp.net/attachments/371759389889003532/747076539790786631/unknown.png |
| 12:55:42 | FromDiscord | <vieru> these are the files |
| 12:56:07 | FromDiscord | <vieru> idk why i have libcrypto.a libcrypto.dll.a |
| 12:56:11 | FromDiscord | <vieru> (edit) 'idk why i have libcrypto.a ... libcrypto.dll.a' => 'idk why i have libcrypto.aand' |
| 12:57:57 | * | FromDiscord quit (Remote host closed the connection) |
| 12:58:22 | * | FromDiscord joined #nim |
| 12:59:08 | FromDiscord | <vieru> my friend sent me this |
| 12:59:10 | FromDiscord | <vieru> https://media.discordapp.net/attachments/371759389889003532/747077470208917534/unknown.png |
| 13:00:28 | FromDiscord | <vieru> leorize: any idea ? |
| 13:09:36 | * | kinkinkijkin joined #nim |
| 13:18:22 | * | luis_ quit (Quit: luis_) |
| 13:18:54 | * | arecacea1 quit (Remote host closed the connection) |
| 13:19:17 | * | arecacea1 joined #nim |
| 13:23:47 | * | luis_ joined #nim |
| 13:32:27 | * | Vladar joined #nim |
| 13:41:36 | * | ja0nz quit (Ping timeout: 265 seconds) |
| 13:42:51 | * | arecacea1 quit (Read error: Connection reset by peer) |
| 13:43:25 | * | arecacea1 joined #nim |
| 13:59:53 | * | krux02 quit (Remote host closed the connection) |
| 14:00:34 | * | krux02 joined #nim |
| 14:02:06 | FromDiscord | <kodkuce> whats difrence betwin int and int literal am trying to clamp some int |
| 14:03:06 | FromDiscord | <kodkuce> or i am writning this clamp wrong |
| 14:03:19 | Yardanico | there shouldn't be any really |
| 14:04:12 | FromDiscord | <kodkuce> oh issue was i supplyed only clamp maximal insted minimal too |
| 14:13:05 | * | luis_ quit (Quit: luis_) |
| 14:32:21 | * | Senketsu joined #nim |
| 14:36:18 | bung | how to represent datetime zero ? |
| 14:37:57 | bung | see https://play.nim-lang.org/#ix=2v0L |
| 14:46:50 | bung | maybe I dont need that |
| 14:54:57 | * | byFate` joined #nim |
| 14:56:45 | FromDiscord | <kodkuce> what |
| 14:58:22 | bung | never mind |
| 14:59:00 | * | byFate` quit (Quit: OMG omG!. The Earth will explode the wicked world. Alien or Amen. https://bit.ly/33tb5lx) |
| 14:59:25 | * | ForumUpdaterBot quit (Remote host closed the connection) |
| 14:59:32 | * | ForumUpdaterBot joined #nim |
| 15:01:40 | shashlick | @vieru why do you have a - after ssl and crypto in the dynlib Override cap |
| 15:01:45 | shashlick | Call |
| 15:02:21 | FromDiscord | <vieru> leorize told me to try so |
| 15:06:04 | FromDiscord | <vieru> here's output of `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl --dynlibOverride:crypto -d:sslVersion:"(" --passL:src/lib/more/openssl/lib/libssl.dll.a --passL: src/lib/more/openssl/lib/libcrypto.dll.a -r tssl.nim` |
| 15:06:14 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v06 |
| 15:06:25 | FromDiscord | <vieru> and here's output of |
| 15:06:27 | FromDiscord | <vieru> `nim c -f -d:ssl -p:. -d:noOpenSSLHacks --dynlibOverride:ssl --dynlibOverride:crypto --passL:src/lib/more/openssl/lib/libssl.dll.a --passL: src/lib/more/openssl/lib/libcrypto.dll.a -r tssl.nim` |
| 15:06:39 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v00 |
| 15:08:35 | shashlick | Don't use dll.a files |
| 15:09:02 | shashlick | Just the libssl.a |
| 15:10:42 | shashlick | The snippet you sent above is asking to link other windows libs |
| 15:10:53 | shashlick | https://hastebin.com/rujanuvoxa.coffeescript |
| 15:11:24 | shashlick | Look for those symbols on Google, it will tell you what lib they are associated with |
| 15:13:04 | shashlick | WSAGetLastError means you need -lWs2_32 |
| 15:13:40 | shashlick | Since your app needs winsock |
| 15:15:53 | FromDiscord | <vieru> ok wait |
| 15:16:00 | FromDiscord | <vieru> imma open my laptop rn |
| 15:18:57 | FromDiscord | <vieru> sent a code paste, see https://play.nim-lang.org/#ix=2v0Z |
| 15:19:04 | FromDiscord | <vieru> ive added -lWs2_32 |
| 15:19:13 | FromDiscord | <vieru> now only these ones show up |
| 15:20:35 | FromDiscord | <vieru> can't figure out what lib is __imp___acrt_iob_func from |
| 15:20:40 | FromDiscord | <vieru> (edit) '__imp___acrt_iob_func from' => '`__imp___acrt_iob_func from`' |
| 15:22:01 | FromDiscord | <vieru> (edit) '`__imp___acrt_iob_func from`' => '`__imp___acrt_iob_func` from' |
| 15:29:45 | ForumUpdaterBot | New thread by Gavr: Does gintro need manually memory manipulations, see https://forum.nim-lang.org/t/6714 |
| 15:33:55 | Zevv | vieru: https://stackoverflow.com/questions/30412951/unresolved-external-symbol-imp-fprintf-and-imp-iob-func-sdl2 ? |
| 15:35:21 | * | luis_ joined #nim |
| 15:35:25 | FromDiscord | <vieru> these are the only libs i can see `SDL2.lib SDL2main.lib legacy_stdio_definitions.lib` |
| 15:35:33 | FromDiscord | <vieru> and when i try to include them |
| 15:35:50 | FromDiscord | <vieru> C:/mingw64/bin/../lib/gcc/x86_64-w64-mingw32/6.3.0/../../../../x86_64-w64-mingw32/bin/ld.exe: cannot find -lSDL2main |
| 15:35:54 | FromDiscord | <vieru> (edit) 'C:/mingw64/bin/../lib/gcc/x86_64-w64-mingw32/6.3.0/../../../../x86_64-w64-mingw32/bin/ld.exe:' => '`C:/mingw64/bin/../lib/gcc/x86_64-w64-mingw32/6.3.0/../../../../x86_64-w64-mingw32/bin/ld.exe:' | '-lSDL2main' => '-lSDL2main`' |
| 15:36:17 | FromDiscord | <vieru> and so on |
| 15:36:47 | FromDiscord | <vieru> and idk how to do this in nim |
| 15:36:49 | FromDiscord | <vieru> `To solve the issue, you can try defining __iob_func() yourself which should return an array containing {*stdin,*stdout,*stderr}.` |
| 15:38:22 | * | luis_ quit (Client Quit) |
| 15:47:05 | shashlick | You are having issues since you are linking vc++ libs with mingw |
| 15:47:11 | disruptek | man the line noise is fierce today. |
| 15:47:30 | disruptek | Zevv: am i supposed to be doing some cps stuff? |
| 15:53:11 | * | thomasross joined #nim |
| 15:53:37 | FromDiscord | <Hearthstone> Is there a speech-to-text lib in Nim? |
| 15:53:59 | Yardanico | I don't think so, you'll have better success using an online API for that |
| 15:54:10 | FromDiscord | <Hearthstone> Oh alright then |
| 16:00:08 | Zevv | disruptek: i dont know who am i to say? |
| 16:03:08 | disruptek | aren't you the one that told me to take the trash out? |
| 16:04:51 | Zevv | i am the last one in the position to tell anyone to do anything these days |
| 16:05:13 | disruptek | are you lying down right now? |
| 16:05:18 | Zevv | sure |
| 16:05:35 | Zevv | i did find out eventually where all those naps came from |
| 16:05:48 | disruptek | you have the 'rona? |
| 16:06:01 | Zevv | no, its all the whisky in my coffee, apparently |
| 16:06:09 | disruptek | ahh, that'll do it. |
| 16:06:14 | Zevv | that'll do it |
| 16:06:41 | disruptek | i figured out what's dumb about concepts. |
| 16:07:00 | Zevv | you cant define a concept by concept |
| 16:07:11 | disruptek | yeah, it'd be funny if it wasn't so sad. |
| 16:07:49 | Zevv | but is it really a *problem* for you? |
| 16:08:10 | disruptek | typing is rarely the problem. |
| 16:08:23 | Zevv | well |
| 16:08:58 | Zevv | for me it has been every now and then. but usually when im trying to push it a bit further then i should need to |
| 16:09:25 | Zevv | cpswise, im not sure what the plan should be now |
| 16:09:44 | Zevv | i was considering a clean room reimplementation with a zero-warn |
| 16:09:52 | disruptek | a wut |
| 16:09:53 | Zevv | and zero-break policy |
| 16:10:13 | Zevv | but not sure if that makes sense to do, because it'll hit the same blocking bugs as we hit now |
| 16:11:05 | disruptek | i have to finish mangling before i fix any other compiler bugs. |
| 16:11:37 | Zevv | I had some other thoughts about things we need to support in the long run. bootstraps should be able to take captured closures into the env I think |
| 16:11:46 | Zevv | captured vars, that is |
| 16:12:21 | disruptek | of course. |
| 16:12:29 | disruptek | i don't think this is a problem. |
| 16:12:34 | Zevv | i was trying to implement the clojure `go` way of making a cps block, but thats typically a proc with no args taking everything from a closure |
| 16:13:00 | Zevv | wow i just saw it. clojure. closure. |
| 16:13:08 | disruptek | lol |
| 16:13:28 | Zevv | sometimes i wonder what the hell i think im doing |
| 16:13:31 | disruptek | java+closure=clojure |
| 16:14:06 | Zevv | if the wrorld wouldnt be better off with me swiping the streeds or brushing donkeys or something |
| 16:14:09 | disruptek | maybe whatelse -> clonure. |
| 16:14:18 | Zevv | nosure |
| 16:14:33 | disruptek | the world would be better off with me showering, but i'm on a roll. |
| 16:15:19 | Zevv | irc hides a lot of the gory details |
| 16:15:20 | disruptek | true. |
| 16:15:50 | Zevv | ill see if i can move over a bit to the diner table |
| 16:16:09 | Zevv | i should be able to handle that |
| 16:16:09 | disruptek | it says a lot that i mistook rotten eggs for my brand of funk miasma. |
| 16:17:15 | * | NimBot joined #nim |
| 16:17:18 | disruptek | right. |
| 16:17:29 | Zevv | drop all the unicode noise and im your man |
| 16:17:47 | disruptek | dude. |
| 16:17:56 | disruptek | you have to turn that shit on manually. |
| 16:18:10 | Zevv | \o/ |
| 16:18:24 | * | Zevv afk, talk talk later |
| 16:18:44 | * | disruptek laments all the donkeys with messy hair. |
| 16:19:44 | FromDiscord | <ache of head> I love this community so much 😆 |
| 16:19:47 | * | ja0nz joined #nim |
| 16:23:56 | * | Senketsu quit (Ping timeout: 258 seconds) |
| 16:26:47 | * | icyphox left #nim ("WeeChat 2.8") |
| 16:42:35 | FromDiscord | <k1tt3hk4t> would I be correct to assume that objects in Nim will always be zero initialized, just like is the case for integers and such? |
| 16:42:42 | disruptek | yes. |
| 16:42:48 | FromDiscord | <k1tt3hk4t> ok nice, thank you |
| 16:42:50 | Yardanico | @k1tt3hk4t yes by default |
| 16:42:54 | Yardanico | you can tweak it of course |
| 16:44:02 | FromGitter | <gogolxdong> Does jester support IPv6 listening? |
| 16:44:10 | FromDiscord | <k1tt3hk4t> another unrelated question: when I implement =destroy and the like, do I need to make those public with * for other files to use them implicitly? |
| 16:44:50 | disruptek | i don't think so, but that's such an excellent question. |
| 16:45:19 | disruptek | you'll find out soon enough. 😉 |
| 16:45:37 | FromDiscord | <k1tt3hk4t> haha good point |
| 16:45:54 | FromDiscord | <k1tt3hk4t> I think I will just to be safe :^P |
| 16:47:08 | Zevv | Yardanico: tweak it? |
| 16:47:24 | Yardanico | Zevv: {.noinit.} |
| 16:47:37 | Yardanico | I don't think you can apply it for objects, but you can apply it for variables |
| 16:47:40 | Zevv | yeah ok, but I cant have it filled with 0x42 instead of 0x00 :) |
| 16:47:46 | Yardanico | 0xDEADBEEF |
| 16:47:53 | disruptek | you can alloc an object w/o alloc0. |
| 16:47:55 | Zevv | and the like |
| 16:48:00 | Zevv | true |
| 16:55:43 | * | hnOsmium0001 joined #nim |
| 16:58:01 | * | superbia1 joined #nim |
| 16:59:43 | leorize | @k1tt3hk4 you don't have to make them public |
| 17:01:13 | * | superbia quit (Ping timeout: 265 seconds) |
| 17:01:42 | FromGitter | <gogolxdong> would Android block the port of application running in Termux? I wrote a httpserver using jester and listen on the same port with App built by Flutter, later works and former doesn't. |
| 17:01:57 | leorize | shashlick: I told them to use a trailing `-` due to a bug in how the compiler do dynlibOverride |
| 17:02:02 | FromDiscord | <k1tt3hk4t> ah cool |
| 17:03:17 | FromDiscord | <k1tt3hk4t> on the topic of destructors, I notice the example page for destructors on the nim-lang website doesn't call the destructor on the element previously occupying the space in the []= function... is that a bug? or is the compiler smart enough to figure out that it has to destroy the data that was previously there somehow |
| 17:03:21 | FromDiscord | <k1tt3hk4t> not exactly clear on how that works |
| 17:04:29 | leorize | it's not a bug |
| 17:05:20 | Zevv | gogolxdong: I don't know, I have a sshd running in termux without problems though. Are you using a high (>1024) port number? |
| 17:05:25 | FromDiscord | <Hearthstone> > would Android block the port of application running in Termux? I wrote a httpserver using jester and listen on the same port with App built by Flutter, later works and former doesn't.↵gogolxdong: No Termux doesn't block the ports. It must either be the configuration or another thing blocking the ports |
| 17:05:34 | FromDiscord | <k1tt3hk4t> ah wait, =sink destroys the destination, ok |
| 17:06:00 | leorize | yea, the point of destructors is to automagically do this stuff :P |
| 17:06:21 | FromDiscord | <k1tt3hk4t> xP |
| 17:06:27 | leorize | also in that example the `=` behavior is defined by the contained type |
| 17:07:18 | Zevv | I've been trying to get started on my "nim memory" part II, which should concentrate on ARC and move semantics and all, but I have problems collecting the outlines |
| 17:07:30 | Zevv | there's a lot of stuff that's related, I don't know how & where to start |
| 17:08:32 | leorize[m] | arc is based on destructors |
| 17:08:34 | disruptek | i would start with the difference between manual memory management and arc. and then you can simply explain how the compiler figures out what to do. |
| 17:08:40 | leorize[m] | so if you cover that first it'd be easy to transistion to arc |
| 17:08:56 | disruptek | this will teach people that there is nothing scary about arc and they don't need to learn anything. |
| 17:09:05 | disruptek | people love learning that they don't need to learn. |
| 17:09:26 | Zevv | should I bring in the old gc? |
| 17:09:34 | disruptek | i wouldn't. |
| 17:09:53 | disruptek | maybe at the end. to show that the performance is better with arc. |
| 17:10:49 | Zevv | also there are still things I'm not sure of myself. How arc & threads should/will play together for example |
| 17:11:00 | Zevv | but I'll figure those out on the fly |
| 17:11:20 | FromGitter | <gogolxdong> yes. Flutter App with 8090 and jester with 8089 |
| 17:11:37 | disruptek | it's simple. |
| 17:11:51 | disruptek | there are no parts of arc that are explicitly thread-relevant. |
| 17:11:55 | leorize | arc uses a shared heap |
| 17:12:02 | leorize | so pointer passing works just as usual |
| 17:12:18 | disruptek | so "moving" memory just involves isolating it so that one thread or another can take over ownership. |
| 17:12:29 | disruptek | it's purely compile-time, conceptually. |
| 17:12:45 | leorize | yep |
| 17:13:01 | Zevv | but who's taking care of the memory barriers / synchonisation |
| 17:13:02 | leorize | I think the new isolated types will be used for cross thread pointer passing |
| 17:13:04 | Zevv | no one is, right |
| 17:13:13 | * | krux02 quit (Remote host closed the connection) |
| 17:13:17 | Zevv | you can't just throw it over the wall |
| 17:13:59 | Zevv | arc will not magically allow shared memory over threads |
| 17:13:59 | * | krux02 joined #nim |
| 17:14:04 | Zevv | there's always a price to pay for that |
| 17:14:06 | disruptek | it doesn't move because it was never in two places at once in the first place. |
| 17:14:19 | disruptek | or, it never exists in two places. |
| 17:14:33 | Zevv | but that's enough |
| 17:14:48 | disruptek | perhaps the key realization that you need to drive home is the concept of single-ownership mutability. |
| 17:15:03 | Zevv | practical example: thread A creates an string. I want thread B to have that and A forget about it |
| 17:15:19 | Zevv | so you're saying we dont do "share nothing" anymore |
| 17:15:29 | Zevv | we will actuall pass the memory ownership between threads? |
| 17:15:36 | disruptek | if thread B frees it, what do you think that requires? |
| 17:15:36 | leorize | yes |
| 17:15:42 | leorize | !rfc isolated |
| 17:15:43 | disbot | https://github.com/nim-lang/RFCs/issues/244 -- 3'isolated' data for Nim 7& 1 more... |
| 17:15:55 | disruptek | it merely requires a free to occur in thread B. when no one is going to use it. |
| 17:15:58 | Zevv | it requires at *least* a mutex or memory barrier to pass the pointer from A to B |
| 17:16:03 | disruptek | why? |
| 17:16:14 | Zevv | How do you share memory over threads? |
| 17:16:18 | Yardanico | it's already shared |
| 17:16:19 | Yardanico | shared heap |
| 17:16:30 | Yardanico | you just "transfer" the ownership from one thread to another |
| 17:16:37 | Zevv | that's a common misconception |
| 17:16:43 | leorize | note that in ARC you don't "share" memory |
| 17:16:46 | Zevv | what if A has a pointer that's still in cache |
| 17:16:52 | Zevv | and B has a stale cache line |
| 17:16:59 | Zevv | that's where locks or barriers come in |
| 17:17:06 | Zevv | NUMA memory is even worse |
| 17:17:17 | disruptek | it's a common misconception that ducks have corkscrew penises. |
| 17:17:29 | Zevv | they don't? |
| 17:17:35 | Zevv | I learned that crabs have 2 |
| 17:17:41 | Zevv | but back to the subject |
| 17:17:46 | federico3 | nim doc -o:./htmldocs --project --index:on src/regex.nim is not including the CSS file - any hint? |
| 17:18:03 | Zevv | there's locality problems when sharing data over threads, no matter how you do your memory management |
| 17:18:15 | Zevv | if you alloc manually it's also there |
| 17:18:31 | leorize | federico3: the compiler likely didn't figure out where the css was |
| 17:18:35 | leorize | we need to make this configurable |
| 17:18:43 | federico3 | it's probably still https://github.com/nim-lang/Nim/issues/14424 :( |
| 17:18:44 | disbot | ➥ Missing documentation on nimdoc.css |
| 17:18:45 | disruptek | Zevv: what you're saying really doesn't make sense; i promise. |
| 17:18:48 | Zevv | I had exactly this discussion with my $team a few weeks ago. It was me against 5 seniors. I felt stupid in the end but man was I right |
| 17:19:08 | leorize | well prime material for nim memory ii then :P |
| 17:19:12 | Zevv | disruptek: you are thread D, I am thread Z. I write something to my "memory", yes? |
| 17:19:13 | disruptek | let's say memory is allocated and its address is JEFF. |
| 17:19:18 | Zevv | you think we share the same memory |
| 17:19:28 | disruptek | no, you are already too far ahead. |
| 17:19:33 | Zevv | but on my core I might have this data available at that address, but you might not |
| 17:19:37 | disruptek | this is all figured out at compile-time. |
| 17:19:46 | Zevv | it might linger in cache. It might be stashed in local NUMA memory. |
| 17:20:20 | Zevv | you *need* to lock it when writing. and you *need* to lock it with the lock when reading |
| 17:20:36 | Yardanico | I don't think you should actually care about the CPU, should you? you should only care about the RAM |
| 17:20:43 | Yardanico | cpu is a kind of "black box" which does it's own thing |
| 17:20:47 | disruptek | actually, he's right about this part. |
| 17:20:51 | leorize | yea |
| 17:20:58 | Zevv | man am I right. I had this fight so hard the other day |
| 17:21:08 | Zevv | I was shouting out crying hot tears |
| 17:21:16 | disruptek | oh dear. |
| 17:21:19 | federico3 | https://en.wikipedia.org/wiki/Memory_barrier |
| 17:21:21 | Zevv | and people going "huh what the hell is this dude rambling about" |
| 17:21:28 | Zevv | "we always did it like this" |
| 17:21:42 | disruptek | well, people have been "doing it wrong" for a long time. |
| 17:21:48 | disruptek | it's kinda what we do. |
| 17:22:01 | disruptek | don't get me started on corkscrew cocks. |
| 17:22:03 | Zevv | I even went to mratsim to ask for solid literature to back my case |
| 17:22:41 | disruptek | races aren't something you should cover in section II. |
| 17:22:49 | Zevv | anyway, my point was: this is not something ARC is involved with |
| 17:23:01 | Zevv | arc does not do atomic ref counts, or does it? |
| 17:23:07 | Zevv | because these are expensive as hell |
| 17:23:11 | disruptek | it's something that could be added into isolate if it isn't there yet. |
| 17:23:17 | disruptek | but i don't know; i haven't followed isolate. |
| 17:23:22 | Zevv | neither have i |
| 17:24:00 | disruptek | anyway, ignore this part of the tutorial. |
| 17:24:18 | leorize | isolate don't allow refs to be passed atm |
| 17:24:29 | Zevv | yeah but people have heart the rumours about arc offering a shared heap. so these questions will have to be addressed somewhere. But sure, I can postpone that |
| 17:25:08 | leorize | you already saw that most don't understand that shared heap is not the end game :P it needs it's own article to address why |
| 17:25:20 | Zevv | haha |
| 17:25:38 | disruptek | right, and it's just one leg of the stool. |
| 17:25:49 | disruptek | don't get me started on corkscrew stools. |
| 17:26:07 | * | superbia1 is now known as superbia |
| 17:27:07 | disruptek | it doesn't do atomic rc because it's not needed. |
| 17:27:15 | leorize | weave does implement memory barriers so we have that as a base to wrap refs |
| 17:27:51 | disruptek | we have memory barries in stdlib, iirc. |
| 17:28:21 | * | kitech1 quit (Quit: ZNC 1.7.5 - https://znc.in) |
| 17:28:23 | disruptek | anyway, it's like 6 lines of code. |
| 17:28:47 | * | kitech1 joined #nim |
| 17:28:52 | Zevv | mratsim is the only one I trust on this :) |
| 17:29:28 | leorize | now I have to go and give all of this an in-depth read because as I expected, I still don't know enough to use threads :P |
| 17:29:32 | Zevv | I made the delibrate decision not to use threads ever again about 10 years ago and stood by it |
| 17:29:38 | disruptek | rude. |
| 17:29:50 | Zevv | well, that was kind of the cool thing with this discussion I had at my customer with all the seniors |
| 17:30:10 | Zevv | I don't use threads because I'm not smart enough and I'm sure i don't understand everything well |
| 17:30:11 | disruptek | threads are like crypto. you shouldn't be afraid to use them unless you know enough to be afraid of your use. |
| 17:30:26 | Zevv | good analogy |
| 17:31:24 | leorize | mratsim will save us with weave :P |
| 17:31:33 | Zevv | nah man, weave is just for the 1% |
| 17:31:39 | Zevv | the other 99% need threads for I/o |
| 17:31:45 | leorize | now I feel like learning rust because they appear to have commoditized threads |
| 17:31:59 | Zevv | that's one of the main reasons to use rust indede |
| 17:32:04 | Zevv | they got your back |
| 17:32:05 | disruptek | i used to think weave would save us but now it's all about arc+cps. |
| 17:32:08 | federico3 | sigh |
| 17:32:10 | Zevv | but you pay a price |
| 17:32:34 | leorize | I have only dabbled around rust pre-1.0 |
| 17:32:44 | leorize | the language still looks ugly as hell but I need to know threads |
| 17:32:45 | disruptek | go is even easier and more expensive. |
| 17:32:55 | Zevv | I was pretty flabbergasted by go. They have this cool concept with channels and share-nothing so the newbies at google would no long write stupid code. But then you find they do _nothing_ to prevent goroutines accessing each others data or globals |
| 17:33:14 | Zevv | so in the end you are just making the same rubbish code again. |
| 17:33:20 | federico3 | A lot of application switch often between being CPU-bound, disk and network-bound. We need abstractions that fit this use-case |
| 17:33:38 | disruptek | hence cps. |
| 17:33:54 | disruptek | cps can move continuations between threads adhoc. |
| 17:34:01 | Zevv | *but* |
| 17:34:15 | Zevv | - let's talk about the implications of that later |
| 17:34:33 | disruptek | well, sure. |
| 17:36:09 | disruptek | fwiw, our csp is gonna suffer the same flaws as go's channels. |
| 17:36:30 | disruptek | in nim's case, these aren't flaws, but you'll be able to make the same criticism in any event. |
| 17:37:10 | Zevv | well, not really |
| 17:37:31 | Zevv | did you try weave? The hoops you have jump through to tell Nim to Shut The Feck Up? |
| 17:37:38 | Zevv | you need to move your data around in ptrs |
| 17:37:51 | disruptek | it's just not what i want. |
| 17:37:52 | Zevv | because nim will go "no no" on you all the time with .gcsafe. |
| 17:37:59 | FromDiscord | <k1tt3hk4t> is there a way to denote that a function returns ownership of a resource without requiring a copy? for instance I want to implement a "pop" function for my data structure, and obviously said data structure will no longer hold any reference to it |
| 17:38:00 | Zevv | that's your protection. You can't just go share data. |
| 17:38:18 | leorize | araq is putting infrastructure in place to let weave deal with refs |
| 17:38:41 | leorize | but mratsim seems to be busy on other stuff atm so we will have to wait |
| 17:39:33 | Zevv | k1tt3hk4t: Nim is often able to infer that, but the safest is to use the "result" variable for returning data should, it implise NVRO |
| 17:39:42 | Zevv | leorize: that is pretty cool |
| 17:39:57 | FromDiscord | <k1tt3hk4t> ah ok, nice |
| 17:39:57 | FromDiscord | <k1tt3hk4t> ty |
| 17:39:58 | leorize | don't quote me on that though, I'm not 100% sure |
| 17:40:29 | Zevv | /implies/ |
| 17:41:48 | Zevv | k1tt3hk4t: if you dare you could also use nim's new "arc" memory management model, there's a lot of infering in place for proper moves there. |
| 17:41:56 | Zevv | --gc:arc |
| 17:42:11 | leorize | --gc:orc if unsure if you have cycles |
| 17:42:35 | FromDiscord | <Recruit_main707> can concepts have fields? |
| 17:42:40 | disruptek | no. |
| 17:42:48 | FromDiscord | <Recruit_main707> ok |
| 17:43:06 | Oddmonger | is it interesting using --gc:none , for perfs ? |
| 17:43:17 | FromDiscord | <Recruit_main707> not really |
| 17:43:27 | FromDiscord | <k1tt3hk4t> friendship ended with cyclic data structures, now array backed graphs are my best friend |
| 17:43:43 | Yardanico | Oddmonger: not with arc :) |
| 17:43:50 | FromDiscord | <Recruit_main707> regions is gc:none on steroids, but arc is already nuts |
| 17:56:18 | * | arecacea1 quit (Remote host closed the connection) |
| 17:57:13 | FromDiscord | <lakmatiol> is there a modulo operator, such that `-1 % 5 = 4` |
| 17:57:27 | FromDiscord | <Recruit_main707> `mod` |
| 17:58:00 | FromDiscord | <Recruit_main707> there is also `mod` for floats in the math module if you need to use it with floats |
| 17:58:21 | * | arecacea1 joined #nim |
| 17:58:23 | FromDiscord | <lakmatiol> that is the remainder, not modulo, at least according to the docs VSC is showing me https://media.discordapp.net/attachments/371759389889003532/747152769718157312/unknown.png |
| 17:59:05 | * | solitudesf quit (Remote host closed the connection) |
| 17:59:09 | FromDiscord | <lakmatiol> I guess I can do `(k + 5) % 5` |
| 17:59:14 | FromDiscord | <lakmatiol> (edit) '%' => 'mod' |
| 17:59:35 | FromDiscord | <lakmatiol> (not in the general case, but in this one it will work) |
| 17:59:54 | * | natrys joined #nim |
| 18:00:42 | leorize | https://nim-lang.org/docs/math.html#floorMod,T,T |
| 18:00:53 | * | solitudesf joined #nim |
| 18:02:21 | FromDiscord | <lakmatiol> thanks |
| 18:02:50 | * | kungtotte quit (Read error: Connection reset by peer) |
| 18:03:08 | * | kungtotte joined #nim |
| 18:06:52 | * | bung quit (Quit: Lost terminal) |
| 18:10:56 | * | solitudesf quit (Remote host closed the connection) |
| 18:11:23 | Zevv | Uhm does 'arc' mean "ar4qs reference counting" or something similar? |
| 18:11:34 | Zevv | and 'orc' is called 'orc' because the 'o' is a circle? |
| 18:11:45 | disruptek | yep. |
| 18:11:53 | Zevv | man that's so bad |
| 18:11:58 | Zevv | I even guessed both right |
| 18:12:13 | Zevv | i probable heard that somewhere |
| 18:16:21 | * | solitudesf joined #nim |
| 18:19:37 | * | endragor quit (Remote host closed the connection) |
| 18:27:34 | * | endragor joined #nim |
| 18:32:18 | * | endragor quit (Ping timeout: 256 seconds) |
| 18:34:59 | Zevv | What are the different ways by with one can "lose" a ref |
| 18:35:07 | disruptek | wut |
| 18:35:08 | Zevv | it can get out of scope. The object of which it was a part can be deleted |
| 18:35:38 | disruptek | refs shouldn't be lost; that's what makes them refs. |
| 18:35:58 | Zevv | ok, not "lost", but dereffed, causing the refcounter to decrease |
| 18:36:07 | disruptek | you called it. |
| 18:36:13 | Zevv | With old-style GC, if I overwrite a pointer with zeros |
| 18:36:22 | Zevv | the GC will scan the stack, find the pointer is gone and free the memory |
| 18:36:37 | Zevv | not something that should happen of course, but stil |
| 18:36:40 | disruptek | de-ref'ing is the tried-and-true method to reduce pesky reference counts. |
| 18:36:41 | FromDiscord | <Clyybber> Zevv: Are you talking about memory synchronization across threads? |
| 18:36:54 | Zevv | clyybber: no I dropped that because everybody is stupid and I am smart |
| 18:37:07 | Zevv | or was it the other way around? dang, I forgot |
| 18:37:21 | Zevv | I got started with my "nim memory" part II |
| 18:37:34 | Zevv | I want to make a readable writeup about arc |
| 18:37:38 | FromDiscord | <Clyybber> oh, i was scrolled up so saw the past |
| 18:37:56 | Zevv | yeah, that was one of the things, I will readlly definately get back to you about that |
| 18:38:07 | Zevv | because that deserves either a section, or a dedicated part III |
| 18:41:02 | FromDiscord | <Clyybber> we probably need to do a barrier on thread move |
| 18:41:35 | FromDiscord | <Clyybber> so that the thread its getting passed to sees the uptodate state of the memory |
| 18:41:37 | disruptek | it needs a part 3, honestly. |
| 18:41:48 | FromDiscord | <Clyybber> thats what you were getting at zevv? |
| 18:42:05 | Zevv | yeah |
| 18:42:49 | Zevv | but what will not be the responsibility of ARC I guess |
| 18:43:00 | FromDiscord | <Clyybber> not sure |
| 18:43:18 | disruptek | it will, afaict. |
| 18:43:22 | Zevv | well, part of the fun you then get is that your destructors will also run in whatever thread happens to trigger the last deref |
| 18:43:26 | FromDiscord | <Clyybber> not of the compiler |
| 18:43:34 | FromDiscord | <Clyybber> but of the stlib |
| 18:43:51 | Zevv | oh probably the stdlib, sure |
| 18:44:06 | FromDiscord | <Clyybber> Zevv now youre jumping ahead of me |
| 18:44:09 | FromDiscord | <Clyybber> what deref |
| 18:44:15 | FromDiscord | <Clyybber> and what destructors |
| 18:44:34 | Zevv | I have a `=destroy` for a thing in thread A |
| 18:44:40 | Zevv | thread A throws the thing over the wall to thread B |
| 18:44:52 | Zevv | And B loses the ref so thing can be destroyed |
| 18:44:53 | FromDiscord | <Clyybber> ok |
| 18:44:53 | Zevv | right |
| 18:44:59 | FromDiscord | <Clyybber> no |
| 18:45:05 | Zevv | good |
| 18:45:05 | FromDiscord | <Clyybber> A doesn't destroy it |
| 18:45:07 | Zevv | explain it to me |
| 18:45:09 | FromDiscord | <Clyybber> only B does |
| 18:45:11 | Zevv | right |
| 18:45:14 | Zevv | that's what I mean |
| 18:45:22 | Zevv | so the =destroy will run in context of thread B |
| 18:45:25 | FromDiscord | <Clyybber> A zeros out its pointer, that disables the destroy |
| 18:45:31 | Zevv | sure |
| 18:46:27 | Zevv | so it's all fun and games, until my =destroy does something funky and I was not propared to have it run from a different thread |
| 18:46:35 | FromDiscord | <Clyybber> well |
| 18:46:38 | disruptek | don't be rude. |
| 18:46:46 | FromDiscord | <Clyybber> ypur destroy shouldn't care |
| 18:46:52 | disruptek | obviously. |
| 18:47:05 | Zevv | no it should not. But this distroy will probably live in a lib someone made and I happen to use |
| 18:47:12 | FromDiscord | <Clyybber> well |
| 18:47:15 | disruptek | so what. |
| 18:47:19 | FromDiscord | <Clyybber> i don't forbid bad code |
| 18:47:32 | Zevv | not complaining, but it's just that some of these things will just still be there |
| 18:47:33 | disruptek | all threads will be on equal footing as far as arc goes. |
| 18:47:33 | FromDiscord | <Clyybber> don't send that stuff over threads then |
| 18:47:48 | FromDiscord | <Clyybber> Zevv: if you have a 'solution'? |
| 18:48:21 | disruptek | free it in the thread in which you want to free it; then that'll be the last write. |
| 18:48:27 | FromDiscord | <Clyybber> you can implement your own destroy that sends it back to the main thread to destroy |
| 18:48:54 | Zevv | no solutions, just wanted to mention that you still have to sit on your hands before you type |
| 18:49:10 | Zevv | it's a shared heap, but you still need all the infrastructure to properly handle that |
| 18:49:21 | Zevv | "of course" you say |
| 18:49:22 | disruptek | let him fix cps first. |
| 18:49:24 | Zevv | "of course" I say |
| 18:49:31 | Zevv | what are you fixing? |
| 18:49:34 | Zevv | cps I hear? |
| 18:49:39 | FromDiscord | <Clyybber> well |
| 18:49:44 | disruptek | #5691 or w/e |
| 18:49:45 | disbot | https://github.com/nim-lang/Nim/issues/5691 -- 3The StmtList processing of template parameters can lead to unexpected errors ; snippet at 12https://play.nim-lang.org/#ix=2tef |
| 18:49:48 | FromDiscord | <Clyybber> currently i don't do anthing |
| 18:49:56 | FromDiscord | <Clyybber> other than cook |
| 18:49:59 | Zevv | clyybber: sure, and that is perefectly fine |
| 18:50:05 | disruptek | no it's not. |
| 18:50:10 | FromDiscord | <Clyybber> lol |
| 18:50:24 | Zevv | it is, because I am on holiday and I'm having a special coffee |
| 18:50:29 | FromDiscord | <Clyybber> good news is I got a good fix |
| 18:50:38 | Zevv | man I got my last good fix years ago |
| 18:50:47 | FromDiscord | <Clyybber> bad news is I need to implement error handling |
| 18:50:51 | FromDiscord | <Clyybber> for shadow scopes |
| 18:51:09 | Zevv | can't you just not make any errors? |
| 18:51:15 | FromDiscord | <Clyybber> sure |
| 18:51:16 | Zevv | just say no to errors |
| 18:51:42 | FromDiscord | <Clyybber> see if you do this echo (var a; 1) |
| 18:51:57 | Zevv | !eval echo (var a; 1) |
| 18:51:58 | NimBot | Compile failed: /usercode/in.nim(1, 12) Error: ':' or '=' expected, but got ';' |
| 18:52:21 | FromDiscord | <Clyybber> nitpicky |
| 18:52:44 | Zevv | what's cooking then? |
| 18:52:50 | FromDiscord | <Clyybber> now if echo is a proc then we don't want a scope there |
| 18:53:03 | FromDiscord | <Clyybber> Zevv: gulash |
| 18:53:08 | Zevv | \o/ |
| 18:53:14 | Zevv | excellent choice |
| 18:53:22 | FromDiscord | <Clyybber> :D |
| 18:53:25 | * | Zevv afk readread |
| 18:53:49 | FromDiscord | <Clyybber> and if there isn't a scope then var a may be a redeclaration |
| 18:53:55 | FromDiscord | <Clyybber> but if echo is a template |
| 18:54:00 | FromDiscord | <Clyybber> with a typed param |
| 18:54:20 | FromDiscord | <Clyybber> then we don't want a redecleration error |
| 18:54:33 | FromDiscord | <Clyybber> so conceptually we want a scope |
| 18:55:11 | FromDiscord | <Clyybber> but we can't decide wether we want a scope or not, because we don't know what we are calling yef |
| 18:55:18 | FromDiscord | <Clyybber> (edit) 'yef' => 'yet' |
| 18:59:36 | * | endragor joined #nim |
| 18:59:41 | * | Zectbumo joined #nim |
| 19:01:56 | FromDiscord | <haxscramper> I made macro proof-of-concept macro library for automatically deriving traits (like in rust) - https://github.com/haxscramper/nimtraits . Can someone give general comments on the idea or implementation or whatever you want to comment on |
| 19:02:38 | FromDiscord | <haxscramper> There is only `Eq`, `Hash` and some basic getter/setter validator - I just want to test the Idea |
| 19:03:29 | kinkinkijkin | alright, i have an es2 helper library that compiles now |
| 19:03:33 | FromDiscord | <haxscramper> I also want to automatically derive constructor for `initT` but don't really know how to declare constructors for case objects so it would be simple to use |
| 19:04:26 | FromDiscord | <haxscramper> If anyone has some 'this would be cool to do' ides I might try implement it |
| 19:04:35 | disruptek | you have to do it manually right now. |
| 19:05:35 | FromDiscord | <haxscramper> Yes, and I don't particularly like it tbh. Especially if I have some kind of field where zero-filled is not a valid value. |
| 19:05:44 | FromDiscord | <haxscramper> Oh, and right now I also support immutable fields |
| 19:05:49 | FromDiscord | <Hearthstone> If Nim had an error handler like Rust |
| 19:05:52 | FromDiscord | <Hearthstone> That'd be cool |
| 19:06:02 | Yardanico | why? |
| 19:06:05 | FromDiscord | <haxscramper> I don't really think it is all that nice |
| 19:06:23 | FromDiscord | <haxscramper> You basically have to write all error handling over and over again, each time |
| 19:09:04 | * | endragor quit (Ping timeout: 256 seconds) |
| 19:10:03 | FromDiscord | <demotomohiro> @treeform |
| 19:10:04 | FromDiscord | <demotomohiro> I wrote this issue about your flippy repo.↵https://github.com/treeform/flippy/issues/26↵Any comment? |
| 19:10:06 | disbot | ➥ Loops in flippy access memory discontinuously ; snippet at 12https://play.nim-lang.org/#ix=2v2l |
| 19:13:55 | FromGitter | <Knaque> I want to make an object where all of its fields should be readable anywhere, but they can only be modified by custom procs. An example of what I mean. (https://hastebin.com/cazetebubo.nim) Is this possible? |
| 19:14:09 | disruptek | sure. |
| 19:15:23 | FromGitter | <Knaque> How might I do that? It's probably something obvious, so I'm prepared to slap myself in the face if that's the case. |
| 19:15:47 | disruptek | how evil are you prepared to get? |
| 19:15:58 | disruptek | how little do you trust your users, i mean? |
| 19:17:40 | FromGitter | <Knaque> Well, I'm basically making a module for personal use, but whenever I do I always want to make it as purpose-built and strict as possible since I often make them public anyway. So yes, I am prepared to be super evil. |
| 19:18:00 | * | bouzu_ quit (Quit: Leaving) |
| 19:18:30 | FromDiscord | <demotomohiro> https://play.nim-lang.org/#ix=2v2p |
| 19:18:37 | FromDiscord | <demotomohiro> @Knaque |
| 19:19:32 | Zevv | clyybber: back, let me readup on your echo (var a; 1) |
| 19:19:41 | FromDiscord | <demotomohiro> `val` is readable in other module but not directly modifable |
| 19:21:09 | FromGitter | <Knaque> It still needs to be modifiable though, just exclusively through my own procs. |
| 19:21:31 | disruptek | define `val=`() |
| 19:22:22 | FromDiscord | <demotomohiro> just define procs that modify the field in same module |
| 19:22:25 | FromDiscord | <demotomohiro> see https://nim-lang.org/docs/manual.html#procedures-properties |
| 19:24:13 | * | ja0nz quit (Remote host closed the connection) |
| 19:26:35 | Zevv | clyybber |
| 19:27:30 | Zevv | how is the gulash |
| 19:29:59 | FromGitter | <Knaque> Thank you disruptek and demotomohiro, that works beautifully. |
| 19:36:24 | kinkinkijkin | trying to build a test application of these gles bindings, but I get undefinded references at the gcc stage, do i have to add something to these bindings to make them explicitly load the library? |
| 19:36:41 | kinkinkijkin | that might not be there, that is |
| 19:36:54 | kinkinkijkin | im using autogenerated bindings from nimio, which is... gone |
| 19:41:08 | Zevv | disruptek: very first part draft: http://zevv.nl/nim-memory/nim-arc.html |
| 19:41:16 | Zevv | I couldn't resist bringing in the GC |
| 19:41:36 | disruptek | ~memory |
| 19:41:37 | disbot | memory: 11The Nim memory model: https://zevv.nl/nim-memory/ |
| 19:41:40 | disruptek | nice. |
| 19:42:16 | disruptek | t will then look at all allocated blocks in memory and try to find which of these are no longer referenced, e.g, there are no longer any pointers referencing the memory block. -- s/referenced/in use/ |
| 19:42:31 | disruptek | s/figures/knows/ |
| 19:42:56 | Zevv | thanks |
| 19:43:38 | disruptek | i would try to remove more words. s/all allocated/allocated/ s/memory block/memory/ /memory lifetime/memory/ etc. |
| 19:44:12 | Zevv | cool |
| 19:44:33 | disruptek | version 1.2 introduced arc. |
| 19:44:38 | Zevv | ah right |
| 19:45:11 | disruptek | okay, this is important: |
| 19:45:21 | disruptek | ARC manages your memory in a totally different way ... |
| 19:45:29 | Zevv | yeah tell me |
| 19:46:15 | disruptek | i think you want to establish garbage collection as "stupid determinations that are easy" and talk about how you could simply do this work manually to achieve better performance, etc. |
| 19:46:33 | disruptek | /then/ you introduce arc as an optimization whereby the compiler does this manual work for you. |
| 19:47:05 | disruptek | this is why i really just want to pitch arc as /not a GC/ -- it's really automatic manual memory management. 😉 |
| 19:47:13 | Zevv | hmm fair enough... let me see how I can fit this in |
| 19:47:21 | disruptek | compile-time manual memory management, let's say. |
| 19:47:33 | Zevv | because on the other hand: most users don't give a crap, they just want stuff that /just works/ |
| 19:47:39 | Zevv | if it talks like gC and walks like a GC |
| 19:47:42 | Zevv | to then it's a GC |
| 19:47:44 | Zevv | them |
| 19:47:49 | disruptek | of course, but we shouldn't tell them it's a gc. |
| 19:48:01 | disruptek | because it really isn't. |
| 19:48:10 | Zevv | --gc:isnt |
| 19:48:21 | FromDiscord | <Rika> --notgc:arc |
| 19:48:24 | Zevv | hehe |
| 19:48:26 | * | endragor joined #nim |
| 19:48:28 | disruptek | nim doesn't need a gc because arc is so clever. |
| 19:48:42 | disruptek | nim doesn't need a gc because the compiler is so clever. |
| 19:49:20 | disruptek | it's kinda like the transition from automatic transmissions to auto-manual double-clutch gearboxes. |
| 19:49:44 | Zevv | the DAF variomatic |
| 19:51:40 | disruptek | s/a counter/an integer/ |
| 19:51:52 | disruptek | s/so it Nim/so Nim/ |
| 19:52:18 | disruptek | i would probably use compiler vs. runtime instead of the more general "Nim". |
| 19:52:33 | disruptek | your audience is programmers, and Nim programmers, to boot. |
| 19:53:10 | disruptek | destructors section looks really good. simple and pretty tight. |
| 19:53:33 | disruptek | s/every single type/each type/ |
| 19:53:48 | disruptek | s/is special/is a/ |
| 19:53:59 | disruptek | saying less means more. |
| 19:54:23 | * | endragor quit (Ping timeout: 258 seconds) |
| 19:54:34 | disruptek | s/provide one for your type/define one/ or similar. |
| 19:54:52 | disruptek | this stuff really is simple. words get in the way. |
| 19:55:26 | Zevv | "With ARC, the compiler got much smarter about memory, so that garbage |
| 19:55:27 | Zevv | collection is no longer needed." |
| 19:57:03 | FromDiscord | <Avatarfighter> the strategy to not have garbage is just to not |
| 19:57:08 | disruptek | A /destructor/ is a `proc` that is called when the reference counter reaches zero. Every type has a destructor. The compiler generates destructors when they are not defined by the programmer. A destructor has the following signature: |
| 19:57:08 | FromDiscord | <Avatarfighter> that is how ARC works 100% |
| 19:57:27 | disruptek | yep. |
| 19:58:14 | disruptek | garbage was a stupid idea, turns out. |
| 19:58:31 | FromDiscord | <Avatarfighter> yeah the idea of having garbage was terrible, so it was commented out |
| 19:58:45 | Zevv | yeah yeah first say it looks really good and then rewrite it |
| 19:59:46 | disruptek | you write it, i'm just suggesting how i'd say less. |
| 19:59:54 | FromDiscord | <Avatarfighter> lmao |
| 19:59:58 | Zevv | I literally copied what you just wrote |
| 20:00:09 | Zevv | apart from some punctuation |
| 20:00:14 | disruptek | it's fine. you get the idea. |
| 20:00:17 | FromDiscord | <Avatarfighter> Are you guys rewriting documentation? |
| 20:01:17 | Zevv | writing |
| 20:01:19 | disruptek | zevv is expanding on his memory article. |
| 20:01:27 | disruptek | ~memory |
| 20:01:28 | disbot | memory: 11The Nim memory model: https://zevv.nl/nim-memory/ |
| 20:01:30 | FromDiscord | <Avatarfighter> Oh that's sick! |
| 20:01:33 | Zevv | there's too many questions floating around about arc and stuff |
| 20:01:45 | Zevv | and araqs doc is really too dense for humans |
| 20:01:48 | FromDiscord | <Avatarfighter> Yeah I can think of so many questions I have |
| 20:01:52 | disruptek | anti-arc-fud campaign. |
| 20:02:01 | FromDiscord | <Avatarfighter> Araq's doc is written in Araq |
| 20:02:12 | Zevv | Avatarfighter: out with them then, now |
| 20:02:16 | Zevv | your questions |
| 20:02:48 | FromDiscord | <Avatarfighter> allow me to read what you guys have said so I can ask questions im sure other people might have |
| 20:03:11 | Zevv | http://zevv.nl/nim-memory/nim-arc.html |
| 20:03:19 | FromDiscord | <Avatarfighter> oh thank you 😄 |
| 20:03:22 | FromDiscord | <Avatarfighter> I was scrolling up |
| 20:03:27 | Zevv | it's just the first snippets |
| 20:03:33 | Zevv | very much work in progress |
| 20:03:49 | FromDiscord | <Avatarfighter> of course dw |
| 20:03:54 | Zevv | I wonder if I should really go all the way talking about `=` and `=sink` already |
| 20:04:08 | Zevv | probbaly, otherwise the whole 'move semantics' thing does not make sense |
| 20:04:11 | FromDiscord | <Avatarfighter> thats actually I a question I have lol |
| 20:04:22 | disruptek | start with = |
| 20:04:49 | disruptek | i mean, this is why i wanted to start with manual memory management. |
| 20:05:02 | disruptek | it gives you an easy way to show the transformation. |
| 20:05:27 | Zevv | yeah I know, but I think going back to manual all the way |
| 20:05:41 | Zevv | that's too far. |
| 20:05:51 | Zevv | people start thinking C or have even no clue what this is about |
| 20:06:02 | FromDiscord | <Avatarfighter> Could you explain what `lent` and `sink` do if you're not writing about that |
| 20:06:15 | Zevv | it's hints to the compiler |
| 20:06:16 | disruptek | so start with the arc version, then show the manual version, then walk through it. |
| 20:06:23 | Zevv | telling it that it is safe not to make a copy |
| 20:06:41 | FromDiscord | <Avatarfighter> that's what lent and sink do? |
| 20:07:30 | FromDiscord | <Avatarfighter> that's really cool actually, I'm reading the destructors docs rn lol |
| 20:07:40 | Zevv | basically yes. The compiler tries to do the right thing without help, it can infer a lot already |
| 20:07:52 | FromDiscord | <Avatarfighter> yeah I see that now, that's super neat haha |
| 20:08:01 | Zevv | you can as a programmer hint the compiler about your intentions so it can do the right thing |
| 20:08:15 | Zevv | it you do not, the worst it will do is copy when it needn't |
| 20:08:47 | FromDiscord | <Avatarfighter> I actually have less questions than I thought tbh |
| 20:09:00 | Zevv | \o/ |
| 20:09:04 | FromDiscord | <Avatarfighter> one thing that might be good to write about is how memory sharing is supposed to work |
| 20:09:16 | Zevv | nooo do'nt go there |
| 20:09:33 | FromDiscord | <Avatarfighter> lmaoooo |
| 20:09:52 | FromDiscord | <Avatarfighter> I need to know Zevv, I'm so interested in how memory sharing is supposed to work with ARC and I'm sure other people would like to know too |
| 20:10:06 | Zevv | yeah that's what I started with today |
| 20:10:08 | disruptek | yeah, tell us. |
| 20:10:16 | disruptek | don't be coy, zevv. |
| 20:10:23 | Zevv | but i'm too stupid for that |
| 20:10:27 | disruptek | awwww |
| 20:10:44 | FromDiscord | <Avatarfighter> im ngl this page should be named something other than destructors bc it talks about so much more than just destructors imo https://nim-lang.org/docs/destructors.html |
| 20:11:11 | Zevv | man I see myself dragging mutexes into destructors end ending up with a deadlock due to priority inversion during a refcount cleanup |
| 20:11:14 | kinkinkijkin | trying to use c2nim on a working c header file and i get "error: token expected: )" on a character that is the token ) |
| 20:11:38 | Zevv | Avatarfighter: well, that's the whole point. It talks about so much more, but otoh, it doesn't |
| 20:12:19 | Zevv | it's all related, but it's hard to see how if you just get that dense doc |
| 20:12:25 | FromDiscord | <Avatarfighter> I see, I just think that its a bit misleading to call that page destructors when it talks about all that stuff ahaha |
| 20:12:33 | * | Zectbumo quit (Remote host closed the connection) |
| 20:13:05 | Zevv | true |
| 20:13:49 | * | Zectbumo joined #nim |
| 20:14:00 | FromDiscord | <Avatarfighter> As I read I see how each thing is seemingly heavily integrated with one another but I only discovered all this info because I've gone through most pages of the documentation except for the ones that sound boring like this |
| 20:14:32 | disruptek | The only way to create a pointer of the ref type is by using new() -- /bzzzzt/ |
| 20:15:16 | FromDiscord | <Avatarfighter> I dont think I've ever used `new()` I don't even know wtf it does I haven't read the documentation bc it sounds boring lmao |
| 20:15:48 | Zevv | Well, ok. Technically nim does the new() for you if you use sugar |
| 20:15:58 | FromDiscord | <Avatarfighter> the sugar lib? |
| 20:16:09 | Zevv | No, let thing = Thing() |
| 20:16:14 | FromDiscord | <Avatarfighter> ohhh |
| 20:16:24 | Zevv | its actually the same as let thing = new Thing |
| 20:16:40 | FromDiscord | <Avatarfighter> `new Thing` is just sugar for `Thing()`? |
| 20:16:49 | FromDiscord | <Avatarfighter> or the equivalent? |
| 20:17:33 | Zevv | I think 'new' was the original |
| 20:17:48 | Zevv | Thing() is handy because you can construct a type with values. Thing(val: 42) |
| 20:17:59 | Zevv | it's a construction shorthand |
| 20:18:18 | FromDiscord | <Avatarfighter> Gosh I really need to learn these terms |
| 20:18:56 | FromDiscord | <Avatarfighter> I feel so stupid when I can't describe what I mean because I don't know the terms I've been programming self taught for 7 years and only just recently started learning the fundamentals of programming in high school lol |
| 20:19:16 | Zevv | Well, it's not always easy. Nim docs are not always very strict on the lingo |
| 20:19:20 | Zevv | you should read the C spec |
| 20:19:26 | FromDiscord | <Avatarfighter> I shall 😛 |
| 20:19:29 | Zevv | definition vs declaration |
| 20:19:55 | Zevv | the nim manual is really not formal enough in these things, but if it were it probably was too dry to read |
| 20:20:24 | FromDiscord | <Avatarfighter> The issue I have with the nim docs is sometimes the docs feel unstructured or are lacking explanation |
| 20:20:43 | FromDiscord | <Avatarfighter> some of them are excellent though |
| 20:20:50 | FromDiscord | <Avatarfighter> others look like I wrote it 💀 |
| 20:20:57 | kinkinkijkin | I have a working install of c2nim and am trying to convert gles2/gl2.h to nim but am met with "gles/gl2.h(395, 50) Error: token expected: )" |
| 20:21:10 | kinkinkijkin | the issue is that 395,50 in that file is ) |
| 20:21:19 | kinkinkijkin | anyone know wtf to do? |
| 20:21:19 | Yardanico | kinkinkijkin: c2nim isn't a 100% automatic tool |
| 20:21:26 | Prestige | Dumb question, are there docs for arrays & sequences? I have seen references in the tutorials but not actual docs for all procedures etc |
| 20:21:27 | Yardanico | you're expected to do some manual edits in gl2.h |
| 20:21:30 | Yardanico | or try nimterop |
| 20:21:41 | Zevv | Avatarfighter: well, theres "the manual", which is really the most important of all. It's pretty long because nim is not a simple language. I've read it dozens of times and still find stuff that I missed |
| 20:21:41 | Yardanico | Prestige: docs for procedures are in system obviously |
| 20:21:57 | kinkinkijkin | im trying to figure out what those edits are in this case but this error makes no sense |
| 20:22:02 | Yardanico | kinkinkijkin: it probably does |
| 20:22:02 | Zevv | for the libs: yes, these docs have varying levels of quality |
| 20:22:09 | Yardanico | you don't need to always limit yourself to that single ) |
| 20:22:13 | Yardanico | look at lines around that |
| 20:22:20 | Prestige | why are they in system? |
| 20:22:22 | Yardanico | check stuff which might seem too complex, some weird C macros |
| 20:22:23 | Yardanico | Prestige: why not? |
| 20:22:24 | FromDiscord | <Avatarfighter> Zevv: I do too, I read the manual often when I take a break from nim. Its just sometimes the docs for the libs are quite lacking e.g: https://nim-lang.org/docs/endians.html |
| 20:22:26 | Yardanico | system is a "base" module |
| 20:22:42 | kinkinkijkin | the only lines before it are #define lines |
| 20:22:57 | Zevv | Avatarfighter: very true. We recently had a little discussion about Tables. That thing as over 130 procs, only for working with tables. |
| 20:23:03 | Prestige | I don't see the connection is all, but thanks for telling me where it is |
| 20:23:06 | Zevv | It's insane, totally unusable and unreadable |
| 20:23:07 | Yardanico | Prestige: wdym? |
| 20:23:11 | kinkinkijkin | then that line is a typedef line, and the character it expects the ) on already is a ), and adding another doesn't even change the error |
| 20:23:13 | Yardanico | system is a _base_ module |
| 20:23:19 | Yardanico | for basic language stuff |
| 20:23:40 | Yardanico | of course (I and others have said this before) we need a way to group procs by their first type in docs |
| 20:23:49 | FromDiscord | <Avatarfighter> Zevv: LOL 130 procs? |
| 20:24:05 | Zevv | Avatarfighter: and personally I find a lot of the examples in the docs very messy because everything has this doAssert() in there. I know why its there but it's just so messy to look at |
| 20:24:20 | Yardanico | Zevv: what's the alternative? |
| 20:24:23 | Zevv | Avatarfighter: yes, and then there's still the iterators and the macros |
| 20:24:36 | Zevv | Yardanico: generated docs vs human readable docs |
| 20:24:41 | Zevv | IT's a lot of work though |
| 20:24:45 | Zevv | and a pain to keep in sync |
| 20:24:51 | Yardanico | Zevv: well nim docs are good enough in a lot of cases |
| 20:24:51 | Prestige | I meant a lexcial correlation |
| 20:24:55 | FromDiscord | <ache of head> I mean, obviously a lack of documentation isn't something good |
| 20:24:57 | FromDiscord | <Avatarfighter> imo the docs need a rewrite, some of them don't feel like they've been touched since before nim 1.0 |
| 20:25:04 | Prestige | lexical* |
| 20:25:07 | Zevv | Yardanico: sure, but also they suck big time in a lot of cases. |
| 20:25:09 | FromDiscord | <ache of head> But for modules like endians you can just kind of make out what things do |
| 20:25:16 | Zevv | I think its still complaint #1 for the newcomers |
| 20:25:22 | FromDiscord | <ache of head> For me the docs for the most part are AMAZING |
| 20:25:31 | Zevv | I can't complain because I hardly betterified any docs myself |
| 20:25:33 | FromDiscord | <ache of head> I think I'm the only person in the world who actually thinks this xD |
| 20:25:37 | FromDiscord | <Avatarfighter> @ache of head I'm sad to say that I spent a whole week trying to figure out how tf im supposed to use Endians lmaooo |
| 20:25:50 | FromDiscord | <ache of head> Ah, sorry 😬 |
| 20:26:02 | Yardanico | well it's not hard |
| 20:26:04 | Yardanico | there's also endians2 |
| 20:26:14 | FromDiscord | <ache of head> I remember I was using the module in one project and I made out what the procs do |
| 20:26:17 | FromDiscord | <ache of head> and which ones I should use |
| 20:26:19 | FromDiscord | <Avatarfighter> Endians2 should replace the stdlib endians imo |
| 20:27:05 | FromDiscord | <Avatarfighter> I've gotten better at nim over time so I can guess more easily how things work/should work but for a new person without much experience I feel like they would cry |
| 20:27:47 | Zevv | Yardanico: and really, I don't know how to do it better, but it doesn't help that for example in tables there is 4 times `[]` in the left column. You can't tell the difference. They are not next to eachother. |
| 20:29:02 | Zevv | only thing I know is that writing docs is just hard and takes so much time. |
| 20:29:05 | Zevv | and it's not fun |
| 20:29:40 | FromDiscord | <Avatarfighter> Yeah, I'm thankful for those who took the time to write the docs and generate them, without them I would be lost. |
| 20:29:50 | FromDiscord | <ache of head> one thing: |
| 20:29:54 | FromDiscord | <ache of head> the docs are **readable** |
| 20:30:09 | FromDiscord | <ache of head> they don't bloat the screen and they make sense design-wise |
| 20:30:27 | Zevv | true that |
| 20:30:36 | FromDiscord | <Avatarfighter> yeah for sure |
| 20:30:41 | kinkinkijkin | agh i cannot find any reason this wouldn't be working in c2nim |
| 20:30:45 | Zevv | I think one big problem often is that the order in most things feels so random |
| 20:30:51 | kinkinkijkin | the error is just too vague |
| 20:30:57 | Zevv | you get 130 procs in a row, but you want the important ones first, probably |
| 20:31:00 | * | hzx joined #nim |
| 20:31:08 | Zevv | or in the top level docs, there's a section 'containers'. Starts with 'critbits' |
| 20:31:12 | Zevv | what the fuck are critbits? |
| 20:31:14 | FromDiscord | <Avatarfighter> Zevv: Especially when the docs were autogenerated and not reordered by a human |
| 20:31:22 | Zevv | The most important one everybody uses is on the bottom: Tables |
| 20:31:25 | Zevv | right |
| 20:31:39 | Zevv | Two times a week someone comes in here on irc looking for dicts or maps |
| 20:31:43 | FromDiscord | <Avatarfighter> OH MY GOD NIM HAS RADIX TREES |
| 20:31:47 | FromDiscord | <Avatarfighter> I NEEDED THIS MONTHS AGO |
| 20:31:59 | FromDiscord | <ache of head> For example (totally not picked on purpose), Rust docs feel very out of place |
| 20:31:59 | FromDiscord | <Avatarfighter> Zevv: ily |
| 20:32:02 | FromDiscord | <ache of head> and unorganized |
| 20:32:04 | FromDiscord | <ache of head> compared to Nim's docs |
| 20:32:31 | FromDiscord | <Avatarfighter> https://nim-lang.org/docs/critbits.html this name is not at all what I thought it is lmao |
| 20:32:42 | Yardanico | well, it's pretty self explanatory though :) |
| 20:32:47 | Zevv | ok, I don't know rust docs well. I had a small affair with Go the other day, and was delighted by the consistency of it all. It was all just at the place where you'd expect it to be |
| 20:32:48 | hzx | Hey! I have used "nimble install tiny_sqlite", and it installed v0.1.2, but it is older, than the recent source from the github page. Can I instruct nimble to download the latest source for packages? |
| 20:32:59 | FromDiscord | <ache of head> > OH MY GOD NIM HAS RADIX TREES↵My man, Nim has CPU relaxing |
| 20:33:00 | FromDiscord | <ache of head> xD |
| 20:33:02 | Zevv | hzx: nimble install <git url |
| 20:33:04 | Yardanico | hzx: nimble install tiny_sqlite@#head |
| 20:33:04 | Zevv | hzx: nimble install <git url> |
| 20:33:09 | FromDiscord | <ache of head> For when you need to "calm your horses" |
| 20:33:10 | Zevv | or that |
| 20:33:27 | hzx | Thanks!!! |
| 20:33:31 | FromDiscord | <Avatarfighter> I needed a Radix a while back because i was being "very efficient" with the way I was storing strings and had like 14GB of strings in memory |
| 20:33:48 | FromDiscord | <ache of head> ofo |
| 20:33:50 | FromDiscord | <ache of head> (edit) 'ofo' => 'oof' |
| 20:34:01 | FromDiscord | <Avatarfighter> my chromebook was very unhappy |
| 20:34:41 | Zevv | wrong tool |
| 20:35:07 | FromDiscord | <lqdev> @ache of head cpuRelax() is used to help the branch predictor in spinlocks |
| 20:35:37 | FromDiscord | <ache of head> What's a spinlock? |
| 20:35:51 | Yardanico | lmgtfy |
| 20:35:55 | Yardanico | https://en.wikipedia.org/wiki/Spinlock :P |
| 20:36:09 | Zevv | ache of head: is it ready? is it ready? is it ready? is it ready yet? is it ready yet? |
| 20:36:09 | FromDiscord | <ache of head> Sorry! |
| 20:36:13 | Zevv | it's like a 5yo |
| 20:36:17 | hzx | the nimble install git-url installs the old one, the name@#head installs the latest one. |
| 20:36:24 | Yardanico | hzx: yes |
| 20:36:31 | Yardanico | nimble installs latest _release_ tag by default |
| 20:36:34 | Yardanico | if tags exist |
| 20:36:37 | FromDiscord | <Avatarfighter> oh my god rtorrent gets spinlocks if you mess up the config file |
| 20:36:38 | FromDiscord | <ache of head> Oh man, this is advanced stuff |
| 20:36:42 | FromDiscord | <ache of head> I get most of it |
| 20:36:44 | Yardanico | not really |
| 20:36:45 | FromDiscord | <ache of head> But still |
| 20:36:50 | Yardanico | it's just while not ready: checkReady() |
| 20:36:53 | Zevv | disruptek: still here? |
| 20:36:54 | Yardanico | mostly |
| 20:37:00 | hzx | yepp, but the latest release tag is an old version in this case |
| 20:37:02 | FromDiscord | <haxscramper> How do I disable `[CommentXIgnored]` warning in `c2nim`? It ingores command flags, nim.cfg, there is nothing in `--help` either |
| 20:37:07 | hzx | but it works now, thanks again |
| 20:37:14 | Yardanico | @haxscramper like other warnings |
| 20:37:20 | Yardanico | ah wait nvm |
| 20:37:30 | FromDiscord | <haxscramper> `c2nim --warning[CommentXIgnored]:off index.h` has no effect |
| 20:37:39 | Yardanico | but does it really matter for c2nim? |
| 20:38:03 | FromDiscord | <haxscramper> what do you mean? |
| 20:38:08 | Zevv | disruptek: wdit if after the refcounts and the destructots we bring it like "well, now you know all you need to know about arc", and then say "...but there's some optimizations we can do" and then bring up the whole move semantics business. |
| 20:38:14 | FromDiscord | <haxscramper> Well, from what I can see it does not matter |
| 20:38:19 | Zevv | because that's basically what it is |
| 20:38:43 | FromDiscord | <haxscramper> Which means there should be some other way of disabling it |
| 20:40:21 | disruptek | yes, but posit it as a question -- what happens when we "copy" a ref and then remove the original instance? |
| 20:40:57 | Zevv | hmm |
| 20:41:07 | disruptek | the only reason move is interesting is because it can cause a destroy, right? |
| 20:41:27 | Zevv | sure |
| 20:42:08 | disruptek | Refs are originally managed in Nim by a garbage collector, or GC in short. -- i would say this more generally, so arc has room to provide "managed pointers" as well as a GC. |
| 20:42:17 | disruptek | bbiab |
| 20:42:31 | kinkinkijkin | what I'm getting from looking into this error right now is that it's expected to just rewrite this entire header from scratch, and since it's an extremely large header (800 lines) that's auto-generated, and I don't know the api it's for very well, the actual solution seems to be "get someone else to do it" |
| 20:44:11 | kinkinkijkin | the only info I've found on it is "use [different thing] instead of [thing in code], closed" |
| 20:53:22 | FromDiscord | <haxscramper> The Fastest solution to disable warnings turned out to be just clone c2nim, find where this warning is printed and add `if msg in {warnCommentXIgnored}:` |
| 20:53:54 | FromDiscord | <haxscramper> Maybe I"m just blind but nothing else worked at all |
| 21:18:54 | * | hzx quit (Quit: Going offline, see ya! (www.adiirc.com)) |
| 21:19:26 | * | guelosk[m] joined #nim |
| 21:23:19 | * | Vladar quit (Quit: Leaving) |
| 21:24:54 | leorize | Zevv: I'm pretty sure the term is "use-after-free" and not the other way around |
| 21:25:09 | * | Jesin quit (Quit: Leaving) |
| 21:29:25 | kinkinkijkin | i just found out what glad is and let me just say, i am very disappointed in myself |
| 21:31:32 | leorize | kinkinkijkin: if you're having issues with c2nim and headers, you should remember that c2nim does not expand C macros by default |
| 21:31:40 | leorize | it attempts to translate them into Nim instead |
| 21:31:59 | kinkinkijkin | hmm |
| 21:32:11 | kinkinkijkin | i dont really need that anymore though, since i just found out what glad is |
| 21:32:25 | leorize | ~c2nim |
| 21:32:26 | disbot | no footnotes for `c2nim`. 🙁 |
| 21:32:30 | kinkinkijkin | glad literally just... generates working bindings for a bunch of different languages, including nim |
| 21:32:38 | kinkinkijkin | glad i found glad |
| 21:32:49 | leorize | ~c2nim is c2nim user manual: https://github.com/nim-lang/c2nim/blob/master/doc/c2nim.rst |
| 21:32:50 | disbot | c2nim: 11c2nim user manual: https://github.com/nim-lang/c2nim/blob/master/doc/c2nim.rst |
| 21:34:05 | Prestige | Neat bot |
| 21:37:39 | FromDiscord | <Recruit_main707> kinkinkijkin: do you have any references for that glad tool? |
| 21:38:11 | kinkinkijkin | i mean there's a website for it and it's FOSS, MIT |
| 21:38:27 | FromDiscord | <Recruit_main707> i cant find it, its name is too generic :p |
| 21:38:40 | kinkinkijkin | google es2 nim glad |
| 21:39:55 | * | PMunch quit (Quit: leaving) |
| 21:39:55 | FromDiscord | <Recruit_main707> nothing still |
| 21:40:59 | kinkinkijkin | https://glad.dav1d.de/ |
| 21:41:50 | FromDiscord | <Recruit_main707> thx 😄 |
| 21:41:55 | * | hnOsmium0001 quit (Ping timeout: 240 seconds) |
| 21:43:38 | * | hnOsmium0001 joined #nim |
| 21:46:06 | * | Zectbumo quit (Remote host closed the connection) |
| 21:47:02 | * | hnOsmium0001 quit (Max SendQ exceeded) |
| 21:47:05 | FromDiscord | <Recruit_main707> what would be faster, constantly allocate and deallocate an array (of 9 floats) on the stack? or rather passing a pointer of that array to a function to update it? |
| 21:48:55 | * | hnOsmium0001 joined #nim |
| 22:00:46 | * | Jesin joined #nim |
| 22:01:25 | * | vicfred quit (Quit: Leaving) |
| 22:02:32 | * | vicfred joined #nim |
| 22:04:08 | * | schurig quit (Ping timeout: 244 seconds) |
| 22:04:54 | * | bung joined #nim |
| 22:05:03 | leorize[m] | it depends |
| 22:05:35 | * | natrys quit (Quit: natrys) |
| 22:05:46 | leorize[m] | if the floats managed to stay in the cache line for the majority of all your calculations, then you might want a heap-alloc array |
| 22:06:01 | leorize[m] | otherwise recalculating might be faster |
| 22:06:33 | disruptek | it's pretty likely that the compiler will be able to keep it in the cpu. |
| 22:07:04 | disruptek | depending on the calculation, though, the calculation might be virtually free. |
| 22:08:24 | leorize[m] | if in doubt just profile everything |
| 22:08:31 | * | tane quit (Quit: Leaving) |
| 22:10:05 | bung | `let f = max(l - 1 - s.find("."),0)` is there better way get number of digits after dot of a float value ? |
| 22:11:05 | FromDiscord | <Recruit_main707> There is no calculation really, just new values |
| 22:14:47 | bung | https://stackoverflow.com/questions/8464131/counting-the-number-of-digits-before-and-after-the-decimal-point-in-c |
| 22:15:03 | bung | seems only string calculation |
| 22:16:00 | bung | ok, thank you! for now leave it as is. |
| 22:18:55 | * | schurig joined #nim |
| 22:25:54 | disruptek | shashlick: what do you think about adding some more path substitutions? any ideas? i am thinking of impl $xdg stuff, maybe. |
| 22:42:47 | * | smitop joined #nim |
| 22:42:58 | FromDiscord | <tomck> sent a code paste, see https://play.nim-lang.org/#ix=2v3l |
| 22:44:35 | * | solitudesf quit (Ping timeout: 258 seconds) |
| 22:49:17 | * | arecacea1 quit (Remote host closed the connection) |
| 22:49:50 | * | arecacea1 joined #nim |
| 22:50:09 | FromDiscord | <tomck> also: with nimble, what's the recommended way to structure examples in your project? |
| 22:50:41 | disruptek | i like runnableExamples but i seem to be in the minority. |
| 22:51:04 | FromDiscord | <dom96> if you mean standalone examples then there are no formal conventions AFAIK |
| 22:51:13 | FromDiscord | <tomck> cool ta |
| 22:58:50 | disruptek | tomck: in your example, i'm not sure what you're trying to achieve. |
| 22:59:00 | disruptek | you don't want `echo x` to print 2? |
| 22:59:41 | disruptek | if you want to run echo at compile-time, put it in a static: block. |
| 23:00:09 | disruptek | the proc gets a {.compileTime.} as does the variable. |
| 23:01:17 | FromGitter | <gogolxdong> @Zevv succeed, Thanks! |
| 23:03:00 | disruptek | tomck: ^ |
| 23:03:30 | FromGitter | <gogolxdong> httpbeast needs a new field of Settings to make something like `let server = newSocket(domain = Domain.AF_INET6)` to support IPv6 listening. It's solved by hard coded though for now. |
| 23:06:12 | FromDiscord | <lqdev> disruptek: runnableExamples is super cool, i have to use it more |
| 23:06:26 | disruptek | it's just hard to debug. |
| 23:06:48 | FromDiscord | <lqdev> but afaik it also runs your code so that's a little problematic for graphics-related stuff |
| 23:06:52 | disruptek | i made a macro somewhere to help output which example it was compiling. |
| 23:07:08 | FromDiscord | <lqdev> i don't want windows poppin' up when i rebuild my docs :( |
| 23:07:30 | disruptek | i'm planning on having my testes emit markdown. |
| 23:08:04 | disruptek | !repo gram |
| 23:08:05 | disbot | https://github.com/disruptek/gram -- 9gram: 11Generic graphs in Nim 15 6⭐ 0🍴 7& 4 more... |
| 23:08:15 | disruptek | good demo of runnableExamples and testes. |
| 23:29:37 | * | lritter quit (Quit: Leaving) |
| 23:39:31 | * | krux02_ joined #nim |
| 23:39:45 | FromDiscord | <dk> You can't easily make VS Code do Nim suggestions in snake_case, right? |
| 23:41:55 | * | krux02 quit (Ping timeout: 240 seconds) |
| 23:42:11 | disruptek | nope, but that's probably because i don't use vscode. |
| 23:43:01 | FromDiscord | <Elegant Beef> Can nimsuggest suggest camelCase idents? |
| 23:44:33 | FromDiscord | <dk> As far as I see it uses the style the ident was written in |
| 23:44:50 | FromDiscord | <Elegant Beef> Yea that's what i figured |
| 23:47:16 | FromDiscord | <Elegant Beef> You could probably change the extension so it inserts underscores wherever there is a case change |
| 23:47:45 | FromDiscord | <Elegant Beef> But snakecasing is bad |
