Documentation Channel for #raku | This channel is logged | Roadmap:
Set by [Coke] on 23 May 2022.
03:10 rf left 12:03 NemokoschKiwi joined 12:13 NemokoschKiwi left
[Coke] 32 tickets left for end of March. 14 are checklists, 4 are Kaiepi's PRs that hopefully require very little work. 13:52
m: say "{(32 / (Date(2023,3,31)} tickets/day"; 13:54
2. it's 2 (and there's a typo in that line.)
and I just closed 2 that apparently lizmat had already completed. :) 14:00
Geth ¦ doc: coke self-assigned encoding list missing in some locations
[Coke] cfa: any outstanding issues on the links PR? 14:09
I addressed at least some of the concerns you had. 14:10
and regarding the testing of doc-website templates - I think the way I did it makes more sense because you have dependencies to run the tests that you would have already had to satisfy on raku/doc - if you instead pull the tests into doc-website via the clone of raku/doc, you may then not be able to run them unless you install the deps (and the tests may require work to deal with being run out of a 14:13
different relative directory)
Geth doc/main: 9b912d27c6 | Coke++ |
doc/main: 838edc4e77 | Coke++ |
Note frequency of site updates isn't immediate.
doc/main: a8e54a3dfa | cfa++ | doc/Language/grammar_tutorial.rakudoc
Fix typo
16:06 cfa joined
cfa morning 16:07
[Coke] la 16:48
16:48 rf joined
cfa [Coke]: links pr looks good from a preliminary run this morning 17:01
lots of failures, though it's possible to address these with some automated rewrites
as for the content tests, i suppose that makes sense. with this in place, though, we'd have each repo cloning the other, which seems messy and coupled. 17:02
i understand that you don't want to duplicate the tests themselves though 17:03
back to the links test: i wonder whether the test itself should skip over 'raku', 'perl' and other tokens that are likely to be false positives. it might also be worthwhile to add a negative look ahead for 'method' so that 'Str method' isn't treated as if it should link to /type/Str 17:07
[Coke] all good points. 17:08
cfa i guess i'm not sure whether it makes sense to merge that links test before addressing most of the failures 17:10
or whether to maintain it in draft, address most of them, then commit 17:11
btw, you've got a rewrite in the pr -- not sure whether that's accidental
[Coke] a rewrite? 17:12
If you mean force push, yes, probably 17:13
cfa no, a tweak to the documentation to satisfy the test
[Coke] Figured the end result of this branch is one commit to add the test, one commit to make it pass. 17:16
cfa one commit to make it pass might make sense, though it'll touch a lot of files 17:17
[Coke] we could back that out and just have the test, I know you had another branch with fixes.
cfa here: Failed 118/416 subtests
[Coke] well, N commits.
cfa (but some of those rewrites are likely incorrect)
188 files changed, 2678 insertions(+), 2678 deletions(-) 17:18
some of those failures are C<Raku> etc. 17:20
and some are rewrites that i missed with these simple patterns
cfa reruns the test in the type-links branch
Geth doc/type-links: b7ab3cd77d | (Will Coleda)++ | doc/Type/Bool.rakudoc
Revert "clean up test descriptions"

This reverts commit a8963e1d620075b044ffcfd482adda5827af3a05.
only for doc/Type/Bool.rakudoc
[Coke] there, now we're back to just the test 17:24
cfa Failed 237/416 subtests 17:26
oh, that's before you did the above
so those rewrites address approx. 119 failures
as for the #index-entry stuff: if the tooling can re-create those older targets, great; if not, i'd suggest we merge this or something similar in the near future. as it stands, a lot of cross-references are broken, which leads to a frustrating user experience 17:34
(and if the targets can be regenerated but not in the near future, it might be worth merging something like this then rolling back once generation is fixed) 17:35
i defer to you on that 17:36
Geth doc/type-links: e52b6c87e5 | (Will Coleda)++ | xt/rakudoc-types.rakutest
Don't complain about perl or raku
[Coke] let's get it working in place, keeping old links working if possible 17:41
ah, see your q to fin. let's see ehat he says 17:42
perl/raku shaves a few failures, not many
cfa agreed re: old links 17:43
[Coke] if the link is something like C<Str> method - This should prob be L<Str method|...>
and so we should complain, but the fix may be not the simple one the test recommends. 17:44
cfa i suppose so 17:46
oh, missed some rewrites 17:48
114 failures here but i'll rerun the test in a few
[Coke] We should add the main suggestion to the styleguide, also, and any others like for methods. 17:54
Checking the text after the link will be harder.
since we're walking the pod structure, and would have to save state... Guessing it's not worth figuring out the coding if we're already down to 114 total. 17:55
would be nifty, though.
cfa hopefully <114 17:56
but i'm not sure how many of these are incorrect rewrites
the changes that i've spotchecked have been correct
okay, Failed 65/416 subtests 18:00
[Coke] that's 65 files, not 65 instances, yes? 18:01
cfa # You failed 65 tests of 416
this is the output of prove
so in doc/Type/independent-routines.rakudoc, # You failed 6 tests of 70 18:02
looks like there're very few files that are failing
oh, no, you're right: 65 files 18:05
[Coke] yah, doing each file as a collection of subtests is much easier to write, but harder to count. 18:15
1/6 is pretty good considering we started with failing so much 18:16
cfa [Coke]: take a look at 18:23
i haven't tried rendering this yet
but you can see the rewrites there
and test them locally
(239 changed files with 2,772 additions and 2,772 deletions) 18:24
this is a child of your type-links branch btw 18:25
[Coke] nice. 18:29
19:06 NemokoschKiwi joined 19:19 NemokoschKiwi left 19:33 cfa left 22:01 raiph joined 23:34 raiph left