timo you should always be able to use "./-" as the path, or an absolute path 00:08
mc2: I would expect we'll be making debian/trixie-backports branches in all the repos under the perl6-team group on salsa, yeah. can you give me more information on why building the moar package no longer works for you? you can see a build on the salsa CI system succeed here: salsa.debian.org/perl6-team/moarvm...bs/9162425 that comes from the debian/sid branch 01:36
[Coke] timo++ 01:49
timo i didn't do anything yet :D 01:59
(that includes going to sleep. which i will do next)
02:27 nativecallable6 left 02:28 releasable6 left, committable6 left, shareable6 left, sourceable6 left, evalable6 left, greppable6 left, coverable6 left, sourceable6 joined, committable6 joined, evalable6 joined, releasable6 joined 02:32 nativecallable6 joined, greppable6 joined, coverable6 joined 02:37 notable6 left, bloatable6 left, unicodable6 left, coverable6 left, evalable6 left, tellable6 left, bisectable6 left, sourceable6 left, releasable6 left, benchable6 left, committable6 left, linkable6 left, greppable6 left, nativecallable6 left, quotable6 left 02:40 evalable6 joined, nativecallable6 joined, bloatable6 joined, unicodable6 joined 02:41 shareable6 joined, benchable6 joined, sourceable6 joined, releasable6 joined, greppable6 joined 02:42 tellable6 joined, linkable6 joined, quotable6 joined, committable6 joined 02:43 coverable6 joined, notable6 joined, bisectable6 joined
jubilatious1_98524 Examples (14) of $*ARGFILES in action: unix.stackexchange.com/search?q=us...4*ARGFILES 04:53
mc2 disbot1: $*ARGFILES also implies (lines), that's why it's important it will respect the cat convention (- is stdin). as timo mentionned: when you want to bypass this, then a path iis used insteed of name and unix people a used to it 06:43
timo: build failed probably because of my own way to build the things .. too late to test: I'll see in the next version. now I git switch trixie-backports from sid branch 06:46
ShimmerFairy I don't see where $*ARGFILES "implies lines", it seems to work like any other IO object. (It also seems a bit curious to me that it's just a synonym for $*IN inside MAIN, but not outside it, but I'm not familiar with $*ARGFILES in general.) 08:06
In general, while tricks like `./-` may be normal in a unix shell, Raku isn't a unix shell, and I think it shouldn't be doing this. If I type `open($file)`, I expect that to open a file named `$file` (assuming it's a legal name ofc); needing to know & remember to handle when that name is `-` feels bad. 08:17
08:18 librasteve_ joined
ShimmerFairy In a similar vein, I don't expect `open("~/file")` to work like on a unix shell, nor do I think it should. 08:20
08:28 finanalyst joined
ab5tract I’m really torn on this issue. I think it’s quite common for users to expect passing ‘-‘ as an argument to a script to indicate stdin 08:28
I’ve been using Linux for 30 years and have never used ‘-‘ for stdout , and if I had ever seen it used or mentioned I have long since forgotten about it 08:29
ShimmerFairy I agree with you on user expectation there, but in my eyes it's the program's responsiblity for interpreting special commandline words like '-', just as much as it is reponsbile for responding to '--option' in the appropriate fashion. 08:34
IMO, the standard library of a programming language should not try to do any of that special stuff, outside of features explicitly designed to make implementing shell utilities easier. 08:36
mc2 ShimmerFairy: raku isn't the shell, raku (as perl before her) is a very good tool to write new unix commands and I as raku user would be very desapointed if it will not. 09:15
think about raku -e 'lines....' 09:18
ShimmerFairy: let's think about it the other way: "-" as a litteral filename is 0.0000000000000000001% of the case and there is an elegant (standard) way to do it (as timo said). - as stdin is much more used. so larry has this moto of "DWIM" 09:21
ShimmerFairy: also: most of the people will be in unix right now so just prepare for the future and honor the past 09:22
patrickb I'm with ShimmerFairy here, don't introduce more magic than necessary. 09:23
Sounds neat now, irreversible regret later. I can already imagine the people shaking their head when reading the documentation of `open` and the subsection "special cases". 09:25
A "-" as a filename will happen from time to time, that's a pretty difficult to debug bug when your bit by it. 09:26
ShimmerFairy Simply put, I don't like the fact that, apparently, in Raku you have to write `open($file eq '-' ?? './-' !! $file)` to make sure it always works (and that's hoping it's the only speical case!), especially since I don't know of another language offhand where you'd be required to do this for the basic I/O features. 09:32
Note that I don't object to Raku making it easy to deal with the - = stdin/out convention in *some* way, just not in the general-purpose I/O functionality. It should be limited to areas where you're explicitly grabbing filenames from the cli, and even then it should be an option. 09:33
10:21 finanalyst left
mc2 patrickb: 30y of perl, much more for the unix toolchain and *noone* regrets it, I guess. in the contrary. 10:52
ShimmerFairy: it should be limited indeed so as I said: CatHandler and lines are the places where "-" as filename would be really desapointing. even python fileinput got it! 10:58
11:02 finanalyst joined
Geth rakudo/main: 6d2fa67e53 | (Elizabeth Mattijsen)++ | 2 files
RakuAST: make =document a directive

As requested in github.com/Raku/RakuDoc-v2/issues/94		 11:14
lizmat finanalyst ^^
11:38 finanalyst left
lizmat hmm.... going up the call stack with callframe we can get at the $!do of the block: do we have a way to get at the package of the $!do ? 12:13
I guess not, eh? timo patrickb ?
s/block/frame 12:14
feels like I'm missing something simple 12:16
13:19 finanalyst joined
[Coke] just got a daily google alert for "perl6 or "perl 6" or raku language"... with a pointer to a "perl 6 fundamentals" article hosted on a gov.mr site. DDG then tells me it's not able to negotiate an SSL connection. 15:46
looks like it might be a lift from a fosdem 2017 presentation based on the snippet in the email 15:48
weird.
16:22 nine left, nine joined
lizmat Now that we have UNIT::<$?LANGUAGE-REVISION> to indicate the language revision level 17:02
it feels that !UNIT_MARKER has become redundant?
ab5tract timo ?
ab5tract Never encountered !UNIT_MARKER before 17:04
lizmat it's how the UNIT:: context is found in the stack 17:07
until nqp::isnull($ctx) || nqp::existskey(nqp::ctxlexpad($ctx), '!UNIT_MARKER') {
$ctx := nqp::ctxouterskipthunks($ctx);
}
17:28 librasteve_ left 18:05 finanalyst left
Geth rakudo/main: 11e48594ac | (Elizabeth Mattijsen)++ | src/core.c/CallFrame.rakumod
Add CallFrame.language-revsion

Returns the language-revision (1 = 6.c, 2 = 6.d, 3 = 6.e) of the invocant, or Nil if none could be found. Mostly intended to be used internally in the core, but may have other uses as well outside of the core, so that's why it's not marked as an implementation detail (just yet). ... (6 more lines)		 18:34
lizmat *revision *sigh*
Geth roast: adc4420a7b | (Elizabeth Mattijsen)++ | 6.c/APPENDICES/A03-older-specs/01-misc.t
Make sure the code that "is_run" executes runs in 6.c as well

To prevent deprecation messages that would appear in 6.d, but not in 6.c		 19:45
rakudo/main: 17b9dc457a | (Elizabeth Mattijsen)++ | src/core.c/Rakudo/Internals.rakumod
Add Rakudo::Internals.cliens-callframe method

This returns the CallFrame object of the first client code that is encountered on the call stack that is *not* part of the core or one of the libraries in lib. If one would consider all of the core and lib to be the same package, functionally similar to what the CLIENT:: PseudoStash offers
To be used in the deprecation of "-".IO and associated		 19:47
lizmat *client *sigh* again :-( 19:48
lizmat shouldn't be committing code today anymore :-)
&
22:44 mc2 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!