|
09:16
sena_kun joined
09:30
sena_kun left
|
|||
| Geth | rakudo/main: c81d9cfe86 | (Elizabeth Mattijsen)++ | 3 files Re-imagine .polymod as an iterator rather than a loop with "gather". Advantages: - is about 1.8x as fast - actual iterator lives in the Rakudo::Iterator class - a value of 1 (aka mod identity) also stops the .polymod, instead of continuing and potentially hanging Fixes #4523 |
09:30 | |
| roast: bd0d3e068a | (Elizabeth Mattijsen)++ | S32-num/polymod.t Add tests for #4523 And a test for non-integer .polymod, which apparently didn't exist yet |
09:31 | ||
| patrickb | PSA: I'm moving this weekend. I'm pretty sure I can't do the binary release today or tomorrow. | 14:21 | |
| lizmat | patrickb: understood, safe moving! | 15:58 | |
| Geth | rakudo/main: da291b5501 | (Elizabeth Mattijsen)++ | src/core.c/Cool.rakumod Simplify Cool.substr-eq candidates Because new-disp is much better at handling these cases and it will make it easier to add new Str.substr-eq candidates |
16:29 | |
| [Coke] | jdv++ | 17:11 | |
| Geth | rakudo/main: c32cb4f6a3 | (Elizabeth Mattijsen)++ | src/core.c/Str.rakumod Allow *-x (aka a Callable) for position in .substr-eq As suggested in #1720 |
17:14 | |
|
18:07
sena_kun joined
19:07
dogbert11 left
|
|||
| ab5tract | I need a little help figuring out how to resolve core-provided libraries to their contents | 19:42 | |
| Here's what I have so far: | |||
| m: my $repo = CompUnit::RepositoryRegistry.repository-for-name("core"); my $test = $repo.candidates("Test").head; dd repo-files => $repo.files("lib/Test.rakumod"); $repo.distribution($test.id).meta<provides><Test><lib/Test.rakumod><file>.IO.e ==> say() | |||
| camelia | :repo-files(Empty) False |
||
| ab5tract | m: my $repo = CompUnit::RepositoryRegistry.repository-for-name("core"); my $test = $repo.candidates("Test").head; dd repo-files => $repo.files("lib/Test.rakumod"); $repo.distribution($test.id).meta<provides><Test><lib/Test.rakumod><file>.IO ==> say() | 19:43 | |
| camelia | :repo-files(Empty) "72A28666F10B17ED5620989C3CAF776808C14EA3".IO |
||
| ab5tract | Sorry, this is a better example: | 19:46 | |
| m: my $repo = CompUnit::RepositoryRegistry.repository-for-name("core"); my $test = $repo.candidates("Test").head; dd repo-files => $repo.files("lib/Test.rakumod"); dd source => $test.meta<source>, :try-read((try $test.read-dist()($test.meta<source>)) // "ERROR") | |||
| camelia | :repo-files(Empty) :source("72A28666F10B17ED5620989C3CAF776808C14EA3") :try-read("ERROR") |
||
| ab5tract | in the stacktrace for `$test.read-dist`, it's clear that it is resolving `source` to an actual path, even though it's only outputting the id here | 19:47 | |
| an actual nonexistent path, rather | |||
| ugexe | it won't be entirely simple because whatever is backing a CUR might not be a file, it could be a socket for instance. so the code is written as such | 19:48 | |
| source is the sha1 part of the path | |||
| it needs to be prefixed with the CURI prefix and /source or some such | |||
| ab5tract | This is an example exception from read-dist: | ||
| Failed to open file /home/camelia/rakudo-m-inst-2/share/perl6/core/dist/72A28666F10B17ED5620989C3CAF776808C14EA3: No such file or directory | 19:49 | ||
| there's nothing that matches the sha1 inside of `./share/perl6` locally either | |||
| ugexe | i guess the first thing that im thinking is i dont know what read-dist is so i assume its an implementation detail as well | ||
| $.prefix.add('sources').add($source-file-name); | 19:51 | ||
| i guess i would have expected it there | |||
| ab5tract | repo.prefix.add("/sources").add($test.meta<source>).IO.slurp.say | 19:52 | |
| this seems to work | |||
| thanks ugexe | |||
| ugexe | my $repo = CompUnit::RepositoryRegistry.repository-for-name("core"); my @files = $repo.installed.map(-> $d { $d.meta<provides>.values.map(-> $v { $v.keys[0] => $d.content($v.keys[0]).IO.absolute }) }).flat; for @files { say $_.raku } | 20:01 | |
| evalable6 | "lib/RakuDoc/To/Text.rakumod" => "/tmp/whatever… | ||
| ugexe, Full output: gist.github.com/57c43b55602ca7f185...ddf236ffa5 | |||
| ugexe | I'd probably write it more like that, although as i mentioned extracting a path from the results of $dist.content(...) isn't specced and won't work for e.g. CompUnit::Repository::Github | 20:02 | |
| also having to do the .keys[0] thing isn't ideal. it is unfortunate we mutate the meta data on install in away that changes the data structure | 20:03 | ||
| (see github.com/rakudo/rakudo/pull/2778 for an attempt I made to fix that) | 20:05 | ||
| ab5tract | www.irccloud.com/pastebin/PxfgDP5C/ | 20:30 | |
| though I guess `CompUnit::*` don't live in core or else don't work this way because they don't resolve | 20:32 | ||
| Let me try with your version | |||
| thanks for the snippet ugexe, that's way nicer than keeping track of an a priori list of modules | 20:34 | ||
| ugexe | my snippet should continue working if we split the core modules into cohesive distributions as well | 20:49 | |
| the distribution stuff only works for modules/distributions, i.e. things under lib/. there is nothing for looking up stuff in src/core* | 20:50 | ||
| you cant `use IO::Path;` for instance | 20:51 | ||
| ab5tract | right, that will require a different approach. but I think it won't be too much work to offer the option of adding a rakudo src directory for looking up files under `src/` | 21:17 | |
| ugexe | what are you trying to do more precisely that would need to know about stuff in src? | 21:18 | |
| ab5tract | just as an option. It's probably not super-helpful for an end user, but it should also be mostly trivial to wire up. | 21:49 | |
| Wiring up the docs is probably more useful, ultimately | |||
|
23:26
sena_kun left
|
|||