Documentation Channel for #raku | This channel is logged | Roadmap:
Set by [Coke] on 23 May 2022.
03:40 rf left
Geth doc/main: 9cfba63d1b | cfa++ | 12 files
Another batch of C< trailing space > changes.
cfa looks like C<>s don't render correctly in tables 04:22
[Coke] Does anything render in tables? 04:32
probably worth a ticket.
Geth doc/main: c046736879 | cfa++ | 10 files
Further C< trailing space > changes.
cfa [Coke]: not sure, and agreed
btw, we have some true positives for C<> spacing on the regexes page
er, false positives 04:34
spacing here is intentional for the table issue 04:39
[Coke] maybe we add a Z<ok> or something?
cfa C<> spacing tests: Failed 14/416 subtests (where at least two are false positives)
[Coke] and update the test to not complain. Or add a tag to the para it's in. 04:40
cfa yeah
[Coke] if you want to take a stab, have fun, else I'll see if I can make it work tomorrow. 04:41
cfa about to call it a night 04:42
[Coke] (y)
cfa making a few more spacing tweaks now 04:43
Geth doc/main: 546f08a665 | cfa++ | 7 files
Further C< trailing space > changes.
cfa nearly there: Failed 7/416 subtests
i'll take another look in the morning 04:47
night! 04:49
04:50 cfa left 14:49 rf joined 15:01 cfa joined
Geth doc/main: d976b4df72 | cfa++ | 6 files
Last batch of trailing space changes.

There are still a couple of test failures in Language/regexes.rakudoc but these are false positives (trailing space is intentional).
doc: coke self-assigned xt/rakudoc-c.rakutest false positives
coke self-unassigned xt/rakudoc-c.rakutest false positives

  cfa++ created pull request #4252: xt/rakudoc-c.rakutest: respect `Z<ignore-code-ws>`
cfa [Coke]: i have a fix for that already
[Coke] cfa... oh, OK. 16:14
cfa see whether you like it, feel free to reject
Geth doc/main: fa65dcc53b | (Will Coleda)++ | 2 files
Rename for some consistency
cfa i riffed on your suggested yesterday
btw leading whitespace isn't being properly tested
[Coke] yah, that looks good! 16:16
cfa it's basically what you suggested
with a hopefully descriptive name
Geth doc/main: 68c898ced5 | cfa++ (committed by Will Coleda) | 2 files
xt/rakudoc-c.rakutest: respect Z<ignore-code-ws>

Z<ignore-code-ws> indicates that the following whitespace check ought to be skipped.
Add a couple of these Z<>s to doc/Language/regexes.rakudoc, pass.
cfa hopefully everything passes now
btw, i checked rendering wasn't affected by the Z<> 16:17
[Coke] cfa++ 16:18
weirdly, that does impact the aspell test 16:20
it thinks the -ws from the Z is part of the text
cfa oh whoops 16:21
feel free to change the comment text
i thought ws would be fine :-/
afk for a bit & 16:22
[Coke] the leading whitespace isn't in the generated C<> objects. 16:24
guessing maybe this is a bug in pod processing?
guess the rendering doesn't show the leading whitespace. 16:26
cfa oh, i see 17:12
Z<ignore-code-ws>C<PostgreSQL is an > is tokenised as ... wsPostgreSQL
a space between the Z<> and C<> there won't affect rendering so that's probably the easiest fix
Geth doc/main: 68f4c76432 | cfa++ | doc/Language/regexes.rakudoc
Add a space between Z<> and C<> to appease aspell

  `Z<ignore-code-ws>C<PostgreSQL is an >` is tokenised as
  ... `wsPostgreSQL` ... during spellcheck.
Adding space between the two codes shouldn't affect rendering.
Addresses (and hopefully closes) #4251.
cfa as for leading spaces, 17:22
doc/Language/regexes.rakudoc:2811:C<SQL>, C< database> that does not leave out any room for matching another
doc/Language/syntax.rakudoc:1009:Wrap an infix operator in C<« »> (or the ASCII equivalent C< << >> >) to
doc/Language/pod.rakudoc:534:To flag text as code and treat it verbatim enclose it in C<C< >>
doc/Language/5to6-perlfunc.rakudoc:436:C<< for %hash.kv -> $k, $v { say "$k: $v" }>>
hmm, the trailing space on syntax.rakudoc:1009 should have been caught by the test 17:23
ah, it's misparsed 17:24
oh, and misrendered: 17:25
i'll fix the pod
Geth doc/main: b23449349b | cfa++ | doc/Language/syntax.rakudoc
Language/syntax.rakudoc: Fix incorrectly formatted C<> code
cfa doc/Language/regexes.rakudoc:2811 is a false positive (whitespace is intentional); pod.rakudoc:534 is a false positive due to my search pattern; 5to6-perlfunc.rakudoc:436 is the only real match there 17:37
Geth doc/main: 25a7d497a6 | cfa++ | doc/Language/5to6-perlfunc.rakudoc
Trim leading space, tidy.
doc/main: 92d2180a36 | cfa++ | doc/Language/5to6-perlfunc.rakudoc
Fix C<> markup
[Coke] cfa: thanks! 17:45
under 460 raku/doc tickets. 18:19
Geth doc/main: 6d47f750d5 | cfa++ | doc/Programs/04-running-raku.rakudoc
Fix broken link (leading slash missing)
doc/main: dee3fbb6f4 | cfa++ | 7 files
Fix a few more broken links
cfa & 18:41
[Coke] added "backlog" milestone for lowest priority. 18:49
coleman [Coke]++ 18:50
Geth ¦ doc: coke self-assigned Definition of infix:<eqv> is not correct 18:55
¦ doc: coke self-unassigned Definition of infix:<eqv> is not correct
doc/main: da8629423e | (Will Coleda)++ | doc/Language/variables.rakudoc
prefer non-html internal links
¦ doc: coke self-assigned Document environment variables for testing 19:08
¦ doc: coke self-assigned Uniform file names in the Languages directory 19:11
¦ doc: coke self-assigned Styling of types 19:27
[Coke] fown to 445 tickets - closed a few stale, transferred a few to doc-website 19:28
i have a test for 19:34
My take: instead of C<tr>, we should use C<L<Str|/type/Str>> - we can complain about the former, and *trust* they convert to the latter 19:35
and the latter won't trigger any issues with the L<> test or the C<> test.
I put in checks for Int/Str/Num and found over 300 instances. 19:36
can expand it to other items that appear in types.
... more like 2700 when I allow for anything in doc/Type to error. :) 19:40
and I wonder if we then also need to verify that anything that looks like L<Hash/type/Hash> is actually immediately wrapped in an C<> 19:42
feedback on this proposal appreciated.
Geth doc/type-links: 969eee6216 | (Will Coleda)++ | xt/rakudoc-c.rakutest
Type names must be links, not just code

part of #2954
[Coke] There's the WIP
fails on 212 files, 2541 subtests fail. 19:47
The ticket also notes that just "Hash" with no C or L is also probably an style error.
... ah, C<L< >> treats the L<> literally. 19:51 19:52
so I guess just L<> is the correct way.
... L<C<>> nests, but C<L<>> does not 19:54
that means the test has to allow these if we're already in an L, thankfully we have prior art there. :) 19:55
20:10 jjatria left 20:11 jjatria joined
[Coke] should be fine to delete old master branch, aye? 20:17
FYI, finanlyst will be less available for a bit, so please consider digging into some of the doc-website tickets if that's your jam. 21:11
Geth doc/type-links: 0fc319dbbc | (Will Coleda)++ | xt/rakudoc-c.rakutest
fix L<C< order, add leading /
[Coke] I may actually put this into a 3rd test. 21:48
the interaction between L & C is a little complicated