| [Coke] | the dd should also have been failing tests. | 00:09 | |
| ... oh, right, the CI on raku/doc is failing, right? | |||
| Geth | doc/coke/fixup-ci: e2926654f9 | (Will Coleda)++ | doc/Type/RakuAST/Name.rakudoc fix POD error |
00:17 | |
| doc: coke++ created pull request #4828: Fix CI failures |
00:20 | ||
| [Coke] | Looks like there is a LOT of dd usage in the new RakuAST pages. | 00:39 | |
| Please avoid dd going forward unless you're specifically trying to showcase dd | |||
| ... ok, I have a test here that is complaining about: | 00:50 | ||
| ... nevermind | |||
| Did we make a change at some point so that a type like a::B had to be a/B in the URL? One of the author tests is complaining about this for all of the new rakuast files, And I'm trying to figure out if the test has been subtly wrong for five years or if something changed on the rendering side | 00:57 | ||
| It's possible the test has always been wrong because most of our classes don't have a namespace | |||
| ah, to link to Distribution::Hash, for example, the expection is that your meta link is /type/Distribution::Hash, not /type/Distribution/Hash | 01:01 | ||
| but the URL is /type/Distribution/Hash | |||
| Not confusing at all. Ok, at least this is a straighforward change. | 01:02 | ||
| Geth | doc/coke/fixup-ci: 5f36fa8d13 | (Will Coleda)++ | doc/Type/RakuAST/Name.rakudoc Pass xt/rakudoc-types.rakutest |
02:04 | |
| doc/coke/fixup-ci: 8827098a56 | (Will Coleda)++ | 13 files Pass xt/rakudoc-types.rakutest |
02:27 | ||
| doc/coke/fixup-ci: 97dbe8e7c4 | (Will Coleda)++ | 8 files Pass t/23-aspell.rakutest |
|||
|
05:39
[Coke] left
07:34
[Coke] joined
|
|||
| [Coke] | ok, there's a lot of 'dd' usage in the new doc files, but instead of naively moving them to be "say"... some of them should be reconsidered. lots of them seem to be methods - in which case they should be changed to the method definition, not showing their usage with a dd or say call | 13:21 | |
| At least now I can see them through the noise of the examples test - thanks for the workaround. I seem to recall there was a reason to do it the convoluted way a decade ago, but that's lost in time. :) | 13:22 | ||
| lizmat | maybe we should add a "syntax-ok" method to Str ? | 13:23 | |
| Geth | doc/coke/fixup-ci: 4075a739a0 | (Will Coleda)++ | t/19-examples-compilation.rakutest Fix test file Remove old, convoluted way to test if something compiles, as this way was broken in the conversion to RakuAST, and required nqp. lizmat++ for the updated, working suggestion |
||
| [Coke] | lizmat: Eh. | ||
| lizmat | or some other HLL way | ||
| [Coke] | This is a step up from the previous nqp-guts way, so I'm happy for now. | 13:27 | |
| just trying to close out "get raku/doc CI green again" right now. | |||
| lizmat | okidoki :-) | ||
| [Coke] | ast seems much more verbose about unhandled exceptions | 13:29 | |
| have another test with a try block with no catch to do anything, on exit, I get a dozen warnings now. | 13:30 | ||
| lizmat | maybe we haz more of them in RakuAST ? | 13:35 | |
| [Coke] | This is all my code that is complaining, though. | 13:36 | |
| added a CATCH { default { .so } } as suggested by the error message does no good. | |||
| s/error/warning/ | |||
|
14:22
Geth left,
Geth joined
|
|||
| [Coke] | ugh, this is annoying. :| | 16:09 | |
| so, you have some references to "Name" in a title, where you are referring to a "RakuAST::Name". I think the heading should use the full class name there. | 16:10 | ||
| (then my "test that the headings are capitalized correctly" script gets harder to make work, because "RakuAST::Name" is two words, not one. :) | |||
| Maybe I'll add a separate pass to grab multi-part class names and whitelist those and deal with RE words last. | 16:11 | ||
|
16:18
disbot left,
disbot joined
|
|||
| lizmat | [Coke]: just tell me what I need to do, and I'll do it :-) | 16:19 | |
| [Coke] | regex for "space or the end of the string" ? ( \s | $ ) not matching end of string | 16:38 | |
| oh, ||, duh | 16:39 | ||
| Would help if I used raku more often. :( | |||
| lizmat | hehe | ||
| [Coke] | ... oof. and when i thought that worked, that's not what I typed. wtf. | 16:41 | |
| Geth | doc/coke/fixup-ci: 475cf076aa | (Will Coleda)++ | t/08-headings.rakutest Fix test Deal with Failures, and allow for valid class names like Foo::Bar to be ignored |
17:13 | |
| doc/coke/fixup-ci: 4bd4316c9b | (Will Coleda)++ | 7 files Pass t/08-headings.rakutest |
|||
| [Coke] | Haven't looked at the guts of these test files in a while. :) | 17:15 | |
| what about macros in the docs? Kill them with fir? | 17:16 | ||
| fire | |||
| lizmat | I guess so ? | 17:17 | |
| hmmm... | |||
| [Coke] | on my branch, make test is down to 't/19-examples-compilation.rakutest' | 17:18 | |
| failures on 36 files. | |||
| Geth | doc/coke/fixup-ci: 664fd790aa | (Will Coleda)++ | doc/Language/experimental.rakudoc Remove notes, examples for macros Keep some text available for search to explain the situation. |
17:22 | |
| [Coke] | github.com/Raku/doc/blob/main/doc/...kudoc#L251 | 17:25 | |
| this used to work as an example, and is now dying with: | |||
| # Cannot compile bind to RakuAST::ApplyListInfix | |||
| Guess we should also have added "test all the examples in raku/doc" before .09 | 17:26 | ||
| How do you want to approach this? want a ticket for each distinct new failure? | 17:27 | ||
| (in rakudo/rakudo) | |||
| lizmat | I guess so ? | ||
| [Coke] | fix could be "update the text and the example in raku/doc" or "actual bug, needs fixing" | ||
| since it could be the former, I'll open the tickets on raku/doc | 17:28 | ||
| this particular example is already sketchy because it's saying "don't do this" and catching the error and showing it to you. | |||
| lizmat | ack | ||
| [Coke] | but, everything else is green - I may just merge this, then the only thing left to worry about is example compilation on the various commits.PRse. | 17:40 | |
| "commits & PRs" | 17:42 | ||
| but I do recommend throwing stuff in a PR, because then we can catch any test failures. | 17:43 | ||
| lizmat | will do so also on doc in the future | ||
| Geth | doc/coke/fixup-ci: 08b368f1d3 | (Will Coleda)++ | util/github-action-test.sh Show the version of raku being used to run the tests in the harness |
19:38 | |
| [Coke] | built a new docker container with 2026.09 | 21:22 | |
| (for the CI) | |||
| Geth | doc/coke/fixup-ci: 38adb8608e | (Will Coleda)++ | README.md TEST commit |
21:28 | |
| [Coke] | ok, all that's left failing is the new examples tests for the rakuast stuff. | 22:17 | |
| and those I think need a structural rewrite to match other type documention. | 22:18 | ||
| Geth | doc/main: 8 commits pushed by (Will Coleda)++ | 22:25 | |
| [Coke] | # Adverb Perl5 not allowed on m | 22:52 | |
| That one's gone now, yes? | 22:53 | ||
| also :P5 ? | |||
| I also see some slang examples failing... | 22:54 | ||
| lizmat | :P5 is gone indeeed | 23:36 | |