28 Mar 2023 | |||
lizmat | the commit message isn't really helpful., and that change goes against what S26 has speculated about | 10:33 | |
but spectests have enshrined that behaviour :-( | 10:34 | ||
it feels wrong to me, as you use formatting information that way? | |||
although I think there's one reason I can think of: after that change, a newline would indicate the boundary between leading and trailing pod | 10:35 | ||
when calling .content | |||
OTOH, you wouldn't be able to see difference between leading and trailing if there was only leading or trailing pod | 10:36 | ||
Geth | rakudo/main: ef05ef1e2b | (Elizabeth Mattijsen)++ | src/core.c/Pod.pm6 Introduce Pod::Declarator class This is basically the Pod::Block::Declarator class, but no longer a subclass of Pod::Block, as Pod declarators are not Pod::Blocks as per S26. This loses two unused attributes per pod declarator. Created a Pod::Block::Declarator class as a subclass of Pod::Declarator to remain source compatible. |
10:43 | |
Xliff | limat: Actually... 20% worse. | 12:10 | |
Regular timings were down. | 12:11 | ||
I'm going to have to build a dedicated timing machine at some point, so I don't have system level tasks colliding with it. I'm running this with all cores, so I'm starting to wonder if that's just too much load on the system and its shaking out as chaotic lockling times. | 12:12 | ||
A dedicated single-user system would answer that question. | |||
[Coke] | m: (8737.780).polymod(60, 60, 24).reverse.say | 13:20 | |
camelia | (0 2 25 37.78) | ||
tonyo | [Coke]++ i'm desperate for a tool like that that just works..one of the only things i miss in go is godoc | 15:21 | |
[Coke] | I think the current user experience is not great: you have to clone the docs, index them with a separate command, and then you can use the installed command line tool. Should be a one stop install with the indexing done and build time. | 15:27 | |
lizmat | Xliff: re timings, I wonder if the parallel timing could benefit from using 1 CPU less than that you actually have | 17:27 | |
Geth | rakudo/main: 9cc7cfe08b | (Elizabeth Mattijsen)++ | t/12-rakuast/sub.rakutest RakuAST: add a bare sub, just to be sure |
18:30 | |
tonyo | just uploaded fez v45, now has a bunch of dist management functions | 19:03 | |
better usage (and command guessing) | |||
Xliff | lizmat: That is a thought. | 19:06 | |
I might want to drop down to physical cores. | 19:07 | ||
tonyo | .tellable fez v45, full dist management | 19:12 | |
hmm, can never remember the right one | |||
.weekly fez v45, full dist management | |||
is there an easy way to view the rendered rakudo docs locally before i submit a PR? | 19:26 | ||
lucs | tonyo: Do you mean rendered from for example a generated README.md? | 20:54 | |
tonyo | i can't get the docs-website to render twice. i'm trying to add some info to the docs.raku.org about dist management | 20:55 | |
pastebin.com/VbtvUKPS | |||
lucs | Maybe the raku-doc channel people can help. | 20:57 | |
Nemokosch | what does "render twice" mean? | 21:15 | |
tonyo | renders the first time (when it does the initial clone) but upon rerunning, it produces the error in the pastebin | 21:17 | |
Nemokosch | sounds new but I haven't built in a while | 21:21 | |
hm, it did work for me but anyway, let's try the "second render" | 22:03 | ||
hm, seems to work | 22:10 | ||
HOWEVER | 22:11 | ||
only now I realize that it could easily be an upgrade of a dependent module or something, that I already have | |||
lol, Doc::TypeGraph:ver<2.2.1.1> doesn't even install without test failures | 22:35 | ||
okay, so here it goes, with up-to-date dependencies... | 22:42 | ||
anyway... it still works | 23:11 | ||
29 Mar 2023 | |||
[Coke] | tonyo: most folks adding content just add the text, you don't have to worry about building and running the site. | 00:27 | |
tonyo | i'll try that | 00:32 | |
i'm adding a big section and am not very goodat pod |