[Coke] can't remember if this is a known issue, just tried to install OpenSSL on windows, "no such file or directory" during the build 13:59
patrickb this sounds like the long paths issue. Can you nopaste the full error? 14:07
[Coke] gist.github.com/coke/dde8dd6554bb4...81925d2bf9 14:20
fired up a project I haven't run on this box since trying a new raku as part of the last release, went to reinstall modules, boom 14:24
patrickb this is on 2024.12? 14:32
ugexe they are on moar-2024.10 15:01
[Coke] I am? ugh. 15:04
I am. will try an update, thanks for noting that. 15:05
(so happy for rakubrew download!) 15:06
I did a JIC 'rakubrew build-zef' which died with "error: unknown option `hard'" 15:08
git version 2.45.2.windows.1 15:09
patrickb ugexe: How did you know?
[Coke] OpenSSL isntalled no issues with rakudo latest, thanks folks. 15:10
"did coke do the basic troubleshooting? probably not." :)
ok, HTTP::UserAgent now complaining. 15:14
"failed to rename" 15:15
(same gist)
ugexe patrickb: one of the lines had something like `... in any statement_control at C:\rakubrew\versions\moar-2024.10\... ...` 15:23
patrickb 😯 15:24
ugexe Not sure why'd OpenSSL would start to install but http::useragent would fail with seemingly the same error 15:27
[Coke]: i'd probably try doing some raku one-liners that rename paths of various lengths to see if you can trigger that error manually 15:29
[Coke]: is it t/001-meta.t specifically that fails and no other tests? 15:36
ugexe my hunch is it would be related to relative paths somehow (which don't work with long paths, so we always need to absolutify paths going to moarvm) and some of the tests look like they pass a relative path to `use lib ...` 15:41
which makes me wonder if this line -- github.com/rakudo/rakudo/blob/3559...kumod#L171 -- needs a `.absolute` on it (it might might also need .resolve as well or whatever resolves . and .., both of which don't work in a long path) 15:42
lizmat wouldn't hurt, would it ?
ugexe maybe here as well github.com/rakudo/rakudo/blob/3559...akumod#L44
well any additional code in the module loading path theoretically increases start-up time 15:43
lizmat but IO::Paths need to be stringified at some point anyway, no?
ugexe well sure, but that doesn't mean we aren't already absolutifying the path elsewhere first or in some situations. if we just add more .absolute calls then we end up going between io::path and strings unneccesarily 15:45
so if it doesn't fix any problem it is detrimental to add it just because 15:46
lizmat ack
anyway, .resolve is the most expensive one
melezhik o/ 17:01
antonantonov: FYI Raku-JavaScript-D3 latest test fails - sparky.sparrowhub.io/report/gh-ant...t-D3/23248 , HTH 17:02
Geth App-Rakubrew/main: 69f92f2ed1 | (Patrick Böker)++ | 2 files
Fix `build-zef` on Windows

It seems `&&` is not a thing in CMD. Running multiple `&&` joined commands in a single `system()` seems to pass the entire thing literally to a single program. In this case git errored out with an "unknown argument --head", presumably because all the args, including the "--head" were passed to the first git.		 19:12
[Coke] nice, thanks 19:17
patrickb Not released yet, but on the way. 19:20
Do you know that the downloaded rakus have a zef included already?
Geth App-Rakubrew/main: 8b3af1dc2b | (Patrick Böker)++ | 4 files
Version 41
 19:34
patrickb It's released. As always, `rakubrew self-upgrade` should do the trick. 19:56

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!