🦋 Welcome to the MAIN() IRC channel of the Raku Programming Language (raku.org). Log available at irclogs.raku.org/raku/live.html . If you're a beginner, you can also check out the #raku-beginner channel!
Set by lizmat on 6 September 2022.
00:22 sorenson left 00:23 sorenson joined 02:19 hulk joined, kylese left 02:28 arkiuat left 02:34 arkiuat joined 02:39 arkiuat left 02:48 arkiuat joined 02:53 arkiuat left 03:01 arkiuat joined 03:15 hulk left, kylese joined 04:32 arkiuat left 04:33 arkiuat joined 04:38 arkiuat left 04:40 arkiuat joined 04:45 arkiuat left 04:50 sibl joined 05:14 arkiuat joined 05:19 arkiuat left 05:24 arkiuat joined 05:28 sibl left 05:29 sibl joined 05:33 arkiuat left 05:45 arkiuat joined 05:50 arkiuat left 05:53 sibl left 06:12 arkiuat joined 06:17 arkiuat left 06:31 sibl joined 06:44 arkiuat joined 06:45 japhb left 06:50 arkiuat left 06:52 japhb joined 07:13 arkiuat joined 07:18 arkiuat left 07:21 arkiuat joined 07:26 arkiuat left 07:40 arkiuat joined 08:51 arkiuat left 08:55 Aedil joined 09:03 arkiuat joined 09:08 arkiuat left 09:14 arkiuat joined 09:19 arkiuat left 09:37 Sgeo left 09:48 arkiuat joined 09:53 arkiuat left 09:56 wayland joined 09:57 wayland76 left 10:15 arkiuat joined 10:20 arkiuat left 10:32 sibl left, abraxxa joined 10:36 abraxxa left 10:37 abraxxa joined 10:48 arkiuat joined 10:53 arkiuat left 11:11 abraxxa left 11:16 arkiuat joined 11:21 arkiuat left
ab5tract [Coke]++ # for consistently pushing the Raku merch options forward 11:41
11:45 arkiuat joined 11:50 arkiuat left 11:56 arkiuat joined 12:00 arkiuat left 12:02 arkiuat joined 12:07 arkiuat left 12:20 arkiuat joined 12:25 arkiuat left, arkiuat joined 12:30 arkiuat left 12:59 arkiuat joined 13:04 arkiuat left 13:19 arkiuat joined 13:24 arkiuat left 13:37 arkiuat joined 13:41 arkiuat left 13:42 arkiuat joined 13:53 sibl joined 13:55 eseyman left, coverable6 left 13:56 greppable6 left, benchable6 left, avar left 13:57 arkiuat left, eseyman joined, avar joined, arkiuat joined 13:58 coverable6 joined 13:59 benchable6 joined, greppable6 joined 14:48 sibl left
[Coke] anyone able to duplicate this aspell error I'm seeing here: github.com/Raku/doc/pull/4789 ? 14:55
It only happens for me in the container.
15:29 gabiruh_ joined, gabiruh left 15:39 arkiuat left 15:52 arkiuat joined
antononcube weekly: rakuforprediction.wordpress.com/20...f-coronel/ 17:32
notable6 antononcube, Noted! (weekly)
ShimmerFairy [Coke]: running `RAKULIB=lib raku t/23-aspell.rakutest`, I just get failed tests from the files with spelling errors. *But*, if I add `LC_ALL=C` to the commandline (temporarily getting rid of my UTF-8 locale), then I get a UTF-8 failure (exact bytes listed change on multiple runs). So I wonder if the errors stem from a non-UTF-8 locale in the test setup. 18:07
[Coke] tried adding LC_ALL=en_US.UTF-8 to the command line, no love. 18:58
no luck on LANG either. 19:08
any pointers on emitting a string using NFKD codepoints instead of NFG? 19:31
nevermind, got it working. 19:32
19:33 Sgeo joined
[Coke] thought maybe generating the dictionary using NFKD might help, but no. 19:42
20:13 arkiuat left 20:14 arkiuat joined 20:23 atcroft left 20:46 Aedil left
ShimmerFairy [Coke]: modifying the test file to spit out the output when I've got the locale set wrong, I notice that the errant 0xE2 always appears to come in the spelling mistake response, e.g. `# Strâ 95` (when interpreted as ISO-8859-1). So far, running `aspell` manually doesn't manifest the same issue, so it's possibly a bug with how Rakudo gets the output? 21:16
I think I'm narrowing down on what the issue could be, actually. 21:24
[Coke]: The problem seems to come down to the ␤ symbol. If you feed text like "Str␤" to an aspell that doesn't understand Unicode (in my tests, using a non-Unicode locale), then it interprets the UTF-8 sequence 53 74 72 E2 90 A4 as "Strâ" followed by the 0x90, which it presumably interprets as a word boundary. "Strâ" is a typo, so it tells you that, and then Raku breaks. 21:35
Looks like aspell can be configured to accept encoding other than the locale default; I wonder if the CI test is using an aspell with an overriding configuration somewhere (hence the locale envvar changes doing nothing in your tests): aspell.net/man-html/The-Options.htm...ng-Options 21:38
There's also an --encoding commandline option, which could be the solution to this issue (insisting aspell interpret input as UTF-8 no matter what the current locale happens to say). Seems to solve the issue for my LC_ALL-based testing, at least. 21:48
21:59 wayland left