|
00:20
Aedil left
00:54
kst joined
01:51
DarthGandalf left
01:52
DarthGandalf joined
02:53
hulk joined
02:54
kylese left
03:05
sibl joined
03:14
sibl left
03:15
hulk left,
kylese joined
03:22
sibl joined
03:40
ShimmerFairy left
04:14
lichtkind__ joined
04:16
lichtkind_ left
05:00
oodani left
05:01
oodani joined
05:48
sibl left
05:50
sibl joined
06:23
lizmat joined
06:28
lizmat left
06:33
Aedil joined
06:41
sibl left
06:42
sibl joined
06:44
Sgeo left
06:45
sibl left,
sibl joined
06:51
sibl left
07:05
sibl joined
07:17
lizmat joined
07:21
abraxxa joined
07:26
lizmat left
07:55
merp joined
08:21
merp left
08:26
lizmat joined
09:39
lizmat left
09:57
abraxxa left,
abraxxa joined
10:07
sibl left
10:09
lizmat joined
10:18
itaipu left,
lizmat left
10:45
lizmat joined
11:10
lizmat left
11:31
librasteve_ left
11:43
lizmat joined
11:45
abraxxa left
11:54
lizmat left
12:09
sibl joined
14:14
Some-body_ joined
14:16
DarthGandalf left
14:22
Some-body_ is now known as DarthGandalf
|
|||
| Voldenet | js also uses `${expr}` | 14:48 | |
| delim.join(list) is not that bad, C# uses string.Join(delim, iterable) instead and it's ugly as hell | |||
| yes, backticks are new magical feature in js, fn`arg` sounds perlish too | 14:50 | ||
| probably shells | 14:51 | ||
| ` If you can trust .. AI:` | 14:52 | ||
| neeext | |||
|
14:53
Sgeo joined
|
|||
| Voldenet | AI is good if you can quickly validate the output | 14:56 | |
|
14:57
sibl left
|
|||
| Voldenet | but history questions are probably risky | 14:58 | |
|
15:40
lizmat joined
15:44
oodani left
15:48
oodani joined
16:17
lizmat_ joined
16:20
lizmat left
16:24
stanrifkin joined
16:25
human_blip left
16:27
human_blip joined
|
|||
| patrickb | Is there a raku-at-fosdem channel or similar? | 16:41 | |
|
17:19
lizmat_ left
|
|||
| Voldenet | hm it may be complicated because of libuv, uv_fs_copyfile is used - code is github.com/MoarVM/MoarVM/blob/6a80...ops.c#L144 | 17:33 | |
| but it's possible that it would work with \\?\ prefix | 17:34 | ||
|
17:45
sivoais left
17:47
sivoais joined
|
|||
| [Coke] | I thought a bit of work had been done to allow long paths by default. | 18:27 | |
| I haven't had issues on windows in a bit. | |||
| ugexe | long paths should generally work in raku code. build-process type things like rakudo makefile or modules that do their own build processes via external tools still likely face those issues though | 18:49 | |
| i.e. installing rakudo to a really long path probably doesn't work | |||
| but using and accessing long paths (such as installing modules) with rakudo should work | 18:50 | ||
| [Coke] | ugexe++ | 18:53 | |
|
18:55
librasteve_ joined
|
|||
| ugexe | github.com/MoarVM/MoarVM/pull/1868 was the final implementation used. github.com/MoarVM/MoarVM/pull/1776 contains most of the discussion on implementation | 18:55 | |
| Generally speaking we do what nodejs does | 18:57 | ||
|
19:58
lizmat joined
|
|||
| Voldenet | Yes, I've tested it and long paths just work | 19:59 | |
| and hacky UNC paths do not | |||
| ugexe | yeah to be clear rakudo itself does not understand the types of paths that actually allow this, i.e. paths prefixed with \\?\ | 20:04 | |
| long paths work because in moarvm we check the path length before doing some path related function, and if its over the limit it gets prefixed to \\?\ | |||
| but like IO::Spec::Win32 doesn't understand paths with \\?\ and thus can't properly use them, for example to create absolute or relative paths | 20:06 | ||
| ideally rakudo would parse such paths correctly so users can explicitly use such paths | 20:07 | ||
|
20:37
Aedil left
21:10
lizmat left
23:14
librasteve_ left
|
|||