00:03 melezhik_ left, melezhik_ joined 00:18 melezhik_ left, melezhik_ joined 00:25 melezhik_ left 00:26 melezhik_ joined 00:36 melezhik_ left, melezhik_ joined 00:42 melezhik_ left 00:43 melezhik_ joined 00:48 melezhik_ left, melezhik_ joined 00:56 melezhik_ left, melezhik_ joined 01:03 melezhik_ left, melezhik_ joined 01:13 melezhik_ left 01:14 melezhik_ joined 01:26 melezhik_ left, melezhik_ joined 01:36 melezhik_ left, melezhik_ joined 01:48 melezhik_ left 01:49 melezhik_ joined 01:54 melezhik_ left, melezhik_ joined 02:04 melezhik_ left, melezhik_ joined 02:09 melezhik_ left, melezhik_ joined 02:16 melezhik_ left, melezhik_ joined 02:22 melezhik_ left, melezhik_ joined 02:29 melezhik_ left, melezhik_ joined 02:36 melezhik_ left 02:37 melezhik_ joined 02:44 melezhik_ left, melezhik_ joined 02:53 melezhik_ left 04:48 melezhik joined
timo github.com/rakudo/rakudo/issues/5762 anyone want to look deeper into this? 11:11
lizmat it's way down on my list 11:20
timo i am in awe of your list 11:42
lizmat heh 11:53
finally spotted the flapper in Pod::To::HTML: 13:13
m: for ^5 { my %h = a => 42, b => 666; say %h.head }
camelia a => 42
a => 42
b => 666
b => 666
a => 42
lizmat newer versions of Template::Mustache will attempt to call a sub/method if it looks like it can 13:14
instead of just returning the field in the hash 13:16
timo aah 13:24
lizmat and Pod::To::HTML testing used a "head" field changing that name fixed the testing issue 13:25
timo was it an intentional change in Mustache? is there a note about this change in behaviour in the change log etc? 13:28
lizmat github.com/raku-community-modules/...1dcd97R431 13:31
thing is, commit was done "after" the last non-community release
and am now starting to wonder whether to revert that 13:32
timo it may be a good idea to exclude methods that come from Any or Cool or something for this logic 13:33
mustache in other languages probably has no syntax for differentiating between method call and attribute access?
lizmat nah, thinking more about this, this feels like a
an attack vector
timo hehe. {{ lmao.EVAL }}
or something like {{ blah.fail }}, plenty of fun methods to try at least for causing some chaos and confusion 13:34
lizmat yes, indeed 13:35
timo it also occurs to me that since this uses can instead of something with cando, and doesn't try with a fallback, any method that requires arguments would be found and called, and immediately die
lizmat that also 13:36
so I think I'm going to remove that feature
timo it does prioritize a hash key over a method call at least
but i'm also not sure if it makes sense to call this an attack vector, don't mustache templates allow arbitrary code in stuff like "if" already? 13:37
lizmat I think the issue is that it's using .defined, rather than :exists 13:39
if an element exists in the hash, it should take that value, regardless of whether it's defined or not? 13:40
timo i don't know how mustache handles undefined values (or rather, how users expect it to behave) 13:41
lizmat neither the only reason I look at it, is because it's a deep dependency of Pod::To::Markdown
timo fair 13:45
15:57 librasteve_ joined 16:12 finanalyst joined 18:06 librasteve_ left 18:08 finanalyst left
lizmat github.com/skaji/mi6/pull/191 Fix for App::Mi6 allowing Raku Community uploads again 18:16
nine: could you make camelia frequent #raku-doc as well ? 18:41
19:36 melezhik_ joined 19:43 melezhik_ left, melezhik_ joined 19:49 melezhik_ left

Powered by Cro and the Raku® Programming Language.

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

Thank you!