🦋 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.
00:00 reportable6 left 00:02 reportable6 joined
ugexe because its set as a my class variable, so its already been set before runtime 00:05
looks like it is a regression from github.com/rakudo/rakudo/commit/06...25cb559e4e 00:06
hythm Thanks 00:12
so this is a genuine bug? if so I can create an issue
ugexe i kinda think so. i would at least expect it to work with `BEGIN %*ENV<RAKUDO_LOG_PRECOMP> = 1; ...` 00:38
00:39 hythm left
ugexe i guess it could also be argued its not a bug. so i dunno 00:40
m: BEGIN %*ENV<RAKUDO_MODULE_DEBUG> = 1; use Test; 00:44
camelia 1 RMD: Attempting 'Test' as a pragma
1 RMD: 'Test' is not a valid pragma
1 RMD: Attempting to load 'Test'
1 RMD: Late loading 'Test'
1 RMD: try-load source at ./rakudo-m-inst-2/share/perl6/core/sources/7446BEA8D62DB1C72975C92A521818D4…
ugexe that works though, which is why i lean towards it might be a bug
although maybe that isn't a good example 00:45
because it might really be
camelia ( no output )
ugexe even though i set it at BEGIN time, the `use Test` also happens then and immediately before 00:46
and in the case of RAKUDO_LOG_PRECOMP, when else should it check the ENV variable but compile time? and if it happens at compile time, its gonna be sequenced before the first character of user code e.g. %*ENV<...> = ... 00:47
i suppose from a purely technical standpoint it is not a bug 00:51
01:00 hythm joined
hythm I see. Thanks for the detailed explanation. 01:00
01:06 codesections left 01:09 Manifest0 left 01:54 wayland joined 02:03 edr left 03:03 linkable6 left, evalable6 left 03:04 evalable6 joined, linkable6 joined 04:04 benchable6 left, squashable6 left, unicodable6 left, tellable6 left, notable6 left, reportable6 left, evalable6 left, committable6 left, linkable6 left, statisfiable6 left, bloatable6 left, releasable6 left, reportable6 joined, unicodable6 joined 04:05 committable6 joined, statisfiable6 joined, releasable6 joined, tellable6 joined, squashable6 joined, linkable6 joined 04:06 benchable6 joined, evalable6 joined 04:07 notable6 joined, bloatable6 joined 04:23 hythm left 04:46 raiph left 05:24 jgaz left 05:41 derpydoo joined 06:00 reportable6 left 06:02 reportable6 joined 06:07 siavash joined 06:26 xinming left 06:28 xinming joined 06:55 MoC joined 07:23 kjp left 07:27 kjp joined 07:30 deoac left 07:44 squashable6 left 07:46 squashable6 joined 07:56 siavash left 08:51 eseyman left 09:01 manu_ joined 09:02 manu_ is now known as eseyman 09:59 Manifest0 joined 10:07 Sgeo left 10:30 derpydoo left 10:41 wayland left 10:42 wayland joined 10:49 MoC left 11:29 lichtkind joined 11:36 teatwo joined 11:37 tea3po joined 11:39 teatime left 11:40 teatwo left 12:00 reportable6 left 12:03 reportable6 joined 12:12 codesections joined 12:40 simcop2387 left, perlbot left 12:42 perlbot joined 12:43 simcop2387 joined, wayland76 joined, wayland left
tbrowder__ hi, i'm runnng into a prob building a new module. it has a script meant for the root user to execute, but a reqular user should be able to run other parts of it. the problem is the script uses routines from a lib/ module and i get permission problems with lib/.precomp. 12:52
for a regular user. 12:53
lizmat no precompilation ?
tbrowder__ there is an existing lib/.precomp created by root, and the a normal user gets access denied when trying to use the script 12:54
12:55 codesections left, thelio joined
tbrowder__ i haven't yet tried manually setting permissions, but i've never seen this behavior before. 12:55
perhaps i should have ensured the precomp was created by the reg user, but a more general solution would be nice to have 12:56
i'm not sure that is possible without an explict root incantation 12:58
chmod 777 12:59
lib 13:00
but using correct syntax 13:03
13:26 guessed joined
guessed hi how can i catch this situation 13:28
m: put Whatever + 1; CATCH { default { .^name.put } }
camelia Use of uninitialized value of type Whatever in numeric context
in block <unit> at <tmp> line 1
guessed with CATCH or with // or with orelse is there a way? 13:29
it is outputting 0 + 1 = 1, i like to detect the inappropriateness and output something different for example 13:30
lizmat m: put Whatever + 1; CONTROL { default { .^name.put } } 13:39
camelia CX::Warn
lizmat warnings are CONTROL exceptions, so you need a CONTROL block instead of a CATCH block
tbrowder__ note the perms prob is only in the dev repo, i don't there should be a prob with zef installation and use 13:43
13:44 thelio left
lizmat could you post a --ll-exception gist of the failure? 13:45
I mean, it should try to write to the first writeable dir
ugexe if you're getting errors from a lib/.precomp then it sounds like you aren't installing the module 13:48
in which case the location of that lib/ folder is probably the only relevant thing 13:49
in your case it sounds like you are doing something that causes it to want to re-precompile 13:50
for instance, changing the repo chain
so i suppose you should tell us A) the full path of the module and how you are adding it to raku (i.e. -I mylib/lib, RAKULIB=mylib/lib, etc) B) how you are invoking raku to use the script as root (like `raku -Ilib bin/myscript.raku) C) how you are invoking raku to use the script as non-root 13:52
i just did `sudo raku -Ilib bin/zef search HTTP`, and then `raku -Ilib bin/zef search HTTP` and it worked. so from what i can tell it should work fine 14:00
although in that case my regular user already has access to lib/, i.e. lib/ is not under /root/ 14:02
hmm, if i do `sudo mkdir foo` then `raku -Ifoo -Ilib bin/zef search HTTP` it gives a bunch of problem messages 14:06
seems related to github.com/rakudo/rakudo/blob/75c6...od#L49-L50 14:07
guessed Lizmat thanks how i can integrate that to conditionally assign value to variable like `my $var = do { Whatever + 1; CONTROL { default { -2 } } }` 14:09
for this to be -2 but this gives Any right now 14:10
lizmat m: my $var is default(-2) 14:13
camelia ( no output )
lizmat not sure what you actually want to achieve 14:14
ugexe strangely in my gist the error occurs with version-matcher and auth-matcher, but never api-matcher 14:15
14:15 raiph joined
guessed i wanted to initalize a variable from either `f()` or `g()`. First i call `&f` and it can give that Warning. If it does, i then call &g and initialize $var with it. `is-default(g())` didn't quite work i guess its compile time or something. i now do `my $var = do { f(); CONTROL { default { } } } // g();` 14:38
so thanks and my other qeuestion is why
m: sub f { $^a, $^b }; f(5, 6)
camelia ( no output )
guessed works but
m: sub f { $^a, $^b }; f.(5 xx 2)
camelia ===SORRY!=== Error while compiling <tmp>
Calling f() will never work with declared signature ($a, $b)
at <tmp>:1
------> sub f { $^a, $^b }; ⏏f.(5 xx 2)
guessed does not
m: sub f { $^a, $^b }; f.(5, 6) 14:39
camelia ===SORRY!=== Error while compiling <tmp>
Calling f() will never work with declared signature ($a, $b)
at <tmp>:1
------> sub f { $^a, $^b }; ⏏f.(5, 6)
guessed &f.(5, 6) works though
is this expected (you can ignore 5 xx 2 one it was a mistake copying)
lizmat f(5.6) // g(7,8) 14:40
guessed but it doesnt fail so it doesnt go to second part 14:41
lizmat my $a = f(5,6) // g(7,8)
guessed Whatever + 100 warns and gives back 100
lizmat so is the warning the problem, or the value ?
guessed warning 14:42
so i now do `my $var = do { f(); CONTROL { default { } } } // g();` after your CONTROL input
warning and the value are both problems
if &f warns, suppress the warning, ignore what it returns, and assign whatever g() gives instead 14:43
lizmat docs.raku.org/language/statement-p...es#quietly
guessed quielty suppresses but doesnt ignore the value returned 14:44
and so g() isn't called
gfldex m: quietly do our $work;¶
camelia ===SORRY!=== Error while compiling <tmp>
Bogus statement
at <tmp>:1
------> quietly do our $work;⏏¶
expecting any of:
gfldex m: quietly do { our $work }; 14:45
camelia ( no output )
guessed im happy with ``my $var = do { f(); CONTROL { default { } } } // g();` not sure if it gets shorter but i don't know of course 14:46
my other question is: 14:47
m: sub f { $^a, $^b }; f.(5, 6)
camelia ===SORRY!=== Error while compiling <tmp>
Calling f() will never work with declared signature ($a, $b)
at <tmp>:1
------> sub f { $^a, $^b }; ⏏f.(5, 6)
guessed this is erroring, is it expected
it doesnt error if &f.(5, 6) is used
15:04 jgaz joined
lizmat sorry, was afk for a bit 15:19
so you only want call g() if there is a warning ? 15:20
guessed yes 15:22
gfldex Is there actually a case where the `.()`-form is required? 15:23
guessed sigilless vars?
lizmat no, I was wondering why you were using &f.(5.6) instead of just f(5,6)
guessed i wasnt using &f.(5, 6) or f(5, 6)
i was using f.(5, 6)
lizmat "it doesnt error if &f.(5, 6) is used" 15:24
OOC, why f.(5,6) ?
gfldex m: sub f { sub g { say 'g' } }; f.();
camelia g
gfldex That is the only case I can thing of. 15:25
guessed m: my \f = { $^a, $^b}; f(3, 4)
camelia ===SORRY!=== Error while compiling <tmp>
Variable '&f' is not declared. Perhaps you forgot a 'sub' if this was
intended to be part of a signature?
at <tmp>:1
------> my \f = { $^a, $^b}; ⏏f(3, 4)
guessed m: my \f = { $^a, $^b}; f.(3, 4).put
camelia 3 4
gfldex And if I see that in a production code-base, I will run away screeming. :->
guessed for the record my f() and g(), and f.(5, 6) were separate questions 15:26
lizmat m: my &f = { $^a, $^b}; f(3, 4).put
camelia 3 4
lizmat guessed: still wondering whether your "only if warning" isn't a symptom you're trying to counter, rather than an underlying issue 15:27
if the underlying issue is undefinedness, then maybe another approach would be better
nemokosch gfldex made a really good case here, that was really hard to read 15:45
f.(5,6) for some &f is basically f()(5, 6) 15:46
this is so confusing I'm not sure it should even be allowed
guessed i see now, i didnt understand when gfldex wrote it 15:47
from documentation it says useful for method calls
nemokosch I don't know, I can't say clearly what seems so irregular about this 15:48
guessed lizmat: im passed a code object. i call it with `Whatever` first. It can give a warning if the code does arithmetic. If it does, i call it with an integer.
nemokosch it's the intersection of two contradictory abstractions
guessed sometimes it really expects Whatever, so it doesnt give a warning 15:49
btw how do I understand it was warning versus failing versus dying? 15:50
15:50 squashable6 left
nemokosch well, the code continued to run and gave no trace 15:50
a Failure is basically an Exception masquerading as a value, it will only explode when you start to treat it like a value 15:52
15:52 squashable6 joined
so in this sense that's not a third option 15:52
related thought: don't ever test for emptiness using .elems 15:53
(unfortunately, the core didn't follow this principle early enough) 15:54
m: say 'So empty I can't even believe it!' if (^Inf).elems
Raku eval Exit code: 1 ===SORRY!=== Error while compiling /home/glot/main.raku Two terms in a row at /home/glot/main.raku:1 ------> say 'So empty I can'⏏t even believe it!' if (^Inf).elems expecting any of: infix infix stopper postfix statement end statement modifier statement modifier loop
nemokosch oops the apostrophe
m: say 'So empty I can't even believe it!' if (^Inf).elems 15:55
Raku eval
nemokosch negations are a hard topic...
15:55 derpydoo joined
the condition was actually for the case when there are elements 15:55
m: say 'So empty I can't even believe it!' unless (^Inf).elems
Raku eval Exit code: 1 ===SORRY!=== Error while compiling /home/glot/main.raku Two terms in a row at /home/glot/main.raku:1 ------> say 'So empty I can'⏏t even believe it!' unless (^Inf).elems expecting any of: infix infix stopper postfix statement end statement modifier statement modifier loop
nemokosch okay I'm gonna unalive myself 💀 15:56
m: sub is-empty($x) { $x.elems }; if is-empty(^Inf) { say 'This is so empty I can't even believe it.'; } 15:57
Raku eval
nemokosch okay, whatever, today I can't think straight
the point is that you would expect something like this to be not empty but if you use a check for the number of elements, the check will be False, as if it had truly 0 elements 15:58
m: (^Inf).elems
Raku eval Exit code: 1 Cannot .elems a lazy list onto a Range in block <unit> at main.raku line 1
nemokosch this is the reason
lizmat and how did this apply to guessed problem? 15:59
nemokosch this does not apply to the problem, this applies to fail vs die vs warn
lizmat ack 16:00
nemokosch .elems produces a Failure that can turn into a False value and do nothing "exceptional"
but in other situations, like the last snippet, it can escalate into an Exception
this is why I said it wasn't really a third option 16:01
lizmat my $var; CATCH { when CX::Warn { $var = g(7,8) } }; $var = f(5,6)
oops, CONTROL of course :-) 16:08
afk& 16:17
16:32 jpn joined 17:13 jpn left 17:14 jpn joined 17:20 deoac joined 17:34 codesections joined 18:05 MoC joined 18:58 MoC left 19:00 codesections left 19:02 Sgeo joined 19:09 hythm joined 19:18 raiph left 19:30 deoac left 19:31 jpn left 19:33 jpn joined 19:34 jpn joined 19:50 jpn left 19:56 jpn joined
tbrowder__ i thought i was clear but the perm problem was during module dev in a local repo, not while zef installing it. i just corrected perms on my local lib dir and all was fine. it just surprised me, but it shouldn't have. i was slinging sudo around like crazy. 20:08
ugexe: what do you think about making "--serial" the default zef install and upgrade commands? 20:10
*default option for install and upgrade commands? 20:11
20:34 jpn left 20:52 guessed left 20:53 hythm left 21:15 ProperNoun left 21:16 ab5tract left 21:17 ab5tract joined 21:26 codesections joined 21:28 codesections left 21:30 codesections joined 21:58 kdon joined 22:22 codesections left 22:30 codesections joined 22:39 jpn joined 22:43 kdon left 22:57 kdon joined
ugexe I dunno, I like how the non serial behavior often shows problems that should be fixed like incorrectly declared depends 23:05
I think I’d be more open to some sort of mechanism that can tell if it needs to rebuild and run tests for subsequent commands
23:07 jpn left
ugexe Which I mention because the only argument for wanting to use serial is to speed up installing dependency chains that aren’t declared fully/properly 23:09
23:09 jpn joined
ugexe It also means you can’t do certain tasks in parallel in the future like install two different distributions at the same time 23:11
23:26 lichtkind left 23:29 Camaleon joined 23:30 AlexDaniel joined
AlexDaniel e: say 42 23:30
evalable6 42
tellable6 2023-03-07T02:40:12Z #raku <jdv> AlexDaniel ok
2023-05-21T15:55:41Z #raku-dev <jdv> AlexDaniel who can fix missing builds on whateverable.6lang.org?
2023-05-22T21:27:12Z #raku-dev <jdv> AlexDaniel around? whateverable builds are not happy.
AlexDaniel hello
okay, last one is… wait when's that 23:31
May 10th but part of 2023.09 release? What am I looking at exactly :o
23:36 tea3po left, tea3po joined 23:40 Camaleon left 23:46 jpn left 23:47 Vyrus left, Vyrus joined 23:48 jpn joined 23:54 teatwo joined 23:57 tea3po left