»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
protium Any idea why I'd be getting this while trying to build moarvm? 01:19
linking libmoar.so
linking moar
./libmoar.so: undefined reference to `mp_set_long'
MasterDuke what os are you using? 01:29
protium Linux 01:31
Gentoo, specifically 01:32
For the most part, on stable branch
MasterDuke hmm, and how did you get moar? 01:38
protium Portage 01:40
The default build environment (like ports for BSD)
2016.10
MasterDuke hmm, not sure. i don't use gentoo, but i'm pretty sure others have successfully built moar on it 01:43
MasterDuke have you tried just cloning it and building manually? 01:44
protium Well, no, but I guess I could do that
In general, I prefer staying within the package management system for obvious reasons (easier to maintain, keep sane env) 01:45
You mean clone it hot off the main branch on github?
MasterDuke yeah
if you don't specify a prefix it'll keep everything in its directory, you don't need to worry about polluting your filesystem 01:46
can at least figure out if it's your system in general or the portage script 01:47
protium Yeah, good point 01:48
protium Alright, will give it a shot 01:48
seatek if I override a method in a child class... is there a way to call the original (non-overridden) method from the child? 04:42
PARENT.method-name doesn't error or warn about anything... but it doesn't call the method either 04:54
MasterDuke m: class A { method a() { say "parent" } }; class B is A { method a() { say "child"; callsame }}; B.a 04:56
camelia rakudo-moar c963ef: OUTPUT«child␤parent␤»
seatek ha! just "callsame"! wow. nice. :) thanks MasterDuke++ ! 05:01
i never would have thought to look for that in the docs under functional programming' 05:02
MasterDuke depends what you want to do though. that only works if you want to call the parent's method from the overridden child's method 05:04
seatek perfect for me! but yeah, i see what you mean. i just wanted to keep the parent method as the "default" and for special cases, override it, but still not have to redefine all the "defaults".. so callsame as the first thing worked perfectly 05:05
(i never can remember if i have to update things in more than 1 place later on) 05:06
never trust yourself is my motto
MasterDuke m: class A { method a() { say "parent a" } }; class B is A { method a() { say "child a" }; method b() { say "child b"; self.^mro[1].a }}; B.b
camelia rakudo-moar c963ef: OUTPUT«child b␤parent a␤»
seatek i knkew there had to be something with the mro 05:07
though why element 1 i have no idea
MasterDuke i would think there's a way with PARENT somehow, but i don't remember what it is 05:08
m: class A { method a() { say "parent a" } }; class B is A { method a() { say "child a" }; method b() { say "child b"; self.^mro }}; B.b
camelia rakudo-moar c963ef: OUTPUT«child b␤»
MasterDuke m: class A { method a() { say "parent a" } }; class B is A { method a() { say "child a" }; method b() { say "child b"; say self.^mro }}; B.b
camelia rakudo-moar c963ef: OUTPUT«child b␤((B) (A) (Any) (Mu))␤»
seatek I tried PARENT::method method() and PARENT. the same
i even tried SUPER ;) 05:09
MasterDuke [1] because it's the parent, [0] would be self, [2] grandparent, etc
seatek hey, that's tidy and makes sense
seatek would your parent at ^mro[1] have a ^mro[1] to their parent, and just no sense of the child? 05:12
MasterDuke m: class A { method a() { say "parent" } }; class B is A { method a() { say "child" }; method b() { say "b"; say self.^mro } }; class C is B { method c() { say self.^mro } }; B.b; C.c 05:13
camelia rakudo-moar c963ef: OUTPUT«b␤((B) (A) (Any) (Mu))␤((C) (B) (A) (Any) (Mu))␤»
seatek that's a lot to stuff into places. is C's stuff refencing B's stuff? It must keep its own scope in each... but compiling the code... this gets complex. i suppose it's ok it took so long 05:20
seatek I started to add documentation for callsame to the secion on overriding methods... but i'm not sure i should. what if people wanted to call with parameters... then they'd have to go the mro route, and that's a little funny 05:34
seatek well, i'm going to put mention of it in there, anyway. i spent a lot of time looking for how to do it and i think it's probably not that uncommon. but mabye people won't want it there because it's a little funny 05:48
dalek c: 9ccd0f2 | seatek++ | doc/Language/classtut.pod6:
How to access overridden parent methods

Maybe should not have this here because it's a little freaky.		 05:49
synopsebot6 Link: doc.perl6.org/language/classtut
dalek c: a15ca0b | gfldex++ | doc/Language/classtut.pod6:
link rather then duplicate content
 08:30
synopsebot6 Link: doc.perl6.org/language/classtut
Ifdog Excuse me but why "prompt" wont response to when in repl in windows? 08:51
seatek gfldex: although an esoteric answer, since it will lead people there from the section on inheritance, that's a much better answer if people can chew for a bit on it. thanks! :) 08:52
Ifdog Wont response to keys .
samcv i love perl6 so much ♥ 08:57
promises are amazing, channels supplies. it makes me so happy 08:58
just came here to say this :P
seatek samcv: i can't wait to start playing with some of that :) 08:59
samcv Proc::Async is great too. though i need to figure out how to get proc::async to capture stderr from rakudo 09:00
i'm running perl6 on moarvm from proc async, i can capture stderr if i do like: note "message"; to print to stderr, but if there's an error in the code i don't see it
but the perl6 bin is a shell file, so maybe i will have different result if i just don't go through that 09:02
nope that didn't fix it hm 09:05
samcv if i run it using the shell and proc async i capture the stderr of what i think is the rakudo compilier and moarvm 09:12
so i guess i'll do that for now
RabidGravy I've got at least one module that has a test that runs a Perl 6 program 09:19
Audio::StreamThing does it to have the icecast client in a different process 09:20
samcv does it use a shell when it runs it? 09:21
RabidGravy let me look
no 09:23
github.com/jonathanstowe/Audio-Str.../030-run.t 09:24
gfldex samcv: see irclog.perlgeek.de/perl6/2016-07-20#i_12870811 09:52
samcv thanks :) 09:54
FROGGS masak: JavaScript really has won :o) - turbo.github.io/ 10:14
we could perhaps do the same with OpenGL or so... 10:15
timotimo OpenCL or cuda, yes 10:33
it even has a bit of OpenCL(?) code in the middle with the "run" function 10:34
RabidGravy on a related note I keep meaning to get an fpga board for a similar purpose 10:56
dalek c: 7093f32 | gfldex++ | doc/Type/Proc/Async.pod6:
fix markup
 11:04
synopsebot6 Link: doc.perl6.org/type/Proc/Async
grondilu how would you deal with anonymous unions with NativeCall? 11:18
e.g. struct Foo { int type; union { char name; int code; }; } 11:19
timotimo grondilu: i'd probably build two versions of Foo and nativecast between them ... 11:25
grondilu completly forgot what nativecast does 11:26
timotimo github.com/timo/SDL2_raw-p6/blob/m...aw.pm#L419 11:27
FROGGS grondilu: you can also stick a union in there... 11:43
m: use NativeCall; class u is repr<CUnion> { has uint8 $.name; has int32 $.code }; class Foo is repr<CStruct> { has int32 $.type; HAS u $.u }; 11:45
camelia ( no output )
FROGGS using nativecast like in the SDL2 example might also be nice, because you'd have a specialized Foo "subclass" then, instead of accessing the union members as $foo.u.name 11:46
gfldex brrt++ # brrt-to-the-future.blogspot.de/2016...ation.html 12:03
tbrowder .tell viki last chance to object to my module breakup/naming scheme at github.com/tbrowder/Misc-Utils-Perl6 12:14
yoleaux tbrowder: I'll pass your message to viki.
tbrowder .tell ilmari last chance to object to my module breakup/naming scheme at github.com/tbrowder/Misc-Utils-Perl6
yoleaux tbrowder: I'll pass your message to ilmari.
RabidGravy okay this windows laptop has been "checking for updates" for over an hour now 12:59
timotimo i've been told windows update has an algorithm in it that's extremely bad when many needed updates pile up 13:00
perhaps like "exponential" bad even
www.wsusoffline.net/ - i'm not sure if this helps
masak top o' the Sunday to you, #perl6 13:07
DrForr Afternoon. 13:10
lizmat masak DrForr timotimo RabidGravy o/ 13:11
RabidGravy Rarr! 13:12
masak .u dinosaur 13:13
yoleaux No characters found
masak (...yet...)
tyil does anyone know how I can add a header to a GET request made with HTTP::UserAgent? I currently have the following snippet termbin.com/bqa9 13:16
tyil but if I simply pass the %headers as the 2nd arg, it cannot resolve the caller 13:17
lizmat RabidGravy might know 13:19
RabidGravy you can do it
though I don't know what "pass the %headers as the 2nd arg, it cannot resolve the caller" means 13:20
an example of it failing would be good 13:22
tyil Use netcat. 13:22
woops 13:23
wrong copy/paste
termbin.com/657y 13:24
timotimo it seems like it wants the header fields as named arguments
so it expects you to flatten your headers hash 13:25
RabidGravy er
you are passing a Hash
you just pass them items as normal named arguments
so something like "$ua.get('foo.com/', Gorilla => 'bananas')" 13:27
that's what the slurpy *%headers does
tyil ah
it doesnt fail to run if I pass it like that
RabidGravy I know
tyil getting a 400 reply now, so I seem to have something to fix still, but now I can continue on that 13:28
thanks
RabidGravy :)
I don't like that AUTOGEN malarkey though, methinks some prototypes in there might be called for 13:29
tyil, if you want finer grained control of the request, it;s make an HTTP::Request object and then pass that to $ua.request 13:30
.get is just a helper for the simple case really
there are also a bunch of helper subs in HTTP::Request::Common for the most usual types of requests 13:32
RabidGravy adds an issue to HUA to remind him to do the prototypes 13:39
tyil with the HTTP::Request and HTTP::UserAgent.request() it works 13:42
thanks a ton :>
tbrowder hi, #perl6 13:48
question of the day ref arrays: is returning a generated container from a sub expensive as compared to filling an input container with the generated data? 13:52
timotimo should be about the same cost 13:53
though if you can re-use the input container a lot, you can get a bit more performance out of the whole thing
but it shouldn't be much
stmuk_ I wonder how many hours of my life are spent waiting for gcc 14:07
FROGGS I guess less than waiting for clang and msvc 14:08
dalek osystem: 3a4ca29 | titsuki++ | META.list:
Add MeCab to ecosystem

See github.com/titsuki/p6-MeCab		 16:24
osystem: 69b5094 | titsuki++ | META.list:
Merge pull request #264 from titsuki/add-mecab

Add MeCab to ecosystem
ajr_ Is anyone around who saw my comments about Rakudo* on Raspberry Pi? 16:26
(Yesterday)
ajr_ Has anyone successfully installed Rakudo* 2016.10 on a Pi? 16:34
(Or given up after failing to do so?) 16:35
DrForr I believe it's been done; I haven't tried however. (I do have an RPi, been meaning to do it.) 16:35
ajr_ Dr. Forr - I've just started trying, and I'm getting errors in the "make" stage. Hoping to learn from others. 16:37
DrForr Well, that would not be me at the moment :/
ajr_ At the moment, my suspicion is that the process is getting killed because it runs out of available memory. 16:39
DrForr Well, that can be solved with ulimit, I do believe. Or at least the OOM killer can be staved off. How much RAM is it taking? 16:41
japhb .tell tadzik Ping re: my Terminal-ANSIColor PR: github.com/tadzik/Terminal-ANSIColor/pull/6 16:53
yoleaux japhb: I'll pass your message to tadzik.
RabidGravy ajr_, I may have done but let me try 16:54
is it by any chance complaining about the compiler?
I seem to recall that the default version of gcc is no good, and you have to install the newest version 16:55
I have unplugged one of the Pis from the network temporarily 'ang on
tadzik japanoise u 16:56
yoleaux 16:53Z <japhb> tadzik: Ping re: my Terminal-ANSIColor PR: github.com/tadzik/Terminal-ANSIColor/pull/6
tadzik arr, keyboard stupid
japhb: thanks, didn't have a moment to test it yet, will doo soon :)
RabidGravy ajr_, yeah I have "gcc (Raspbian 4.8.2-21~rpi3rpi1) 4.8.2" I believe the default is something 3ish 16:57
DrForr m: class Foo::Bar is Foo { }; class Foo { };
camelia rakudo-moar 736ab1: OUTPUT«5===SORRY!5=== Error while compiling <tmp>␤'Foo::Bar' cannot inherit from 'Foo' because it is unknown.␤at <tmp>:1␤»
RabidGravy but just building to check that nothing gone broken in the meantime anyway 16:58
DrForr Interesting, my error must be to do with roles then.
RabidGravy ajr_, so if you haven't done so already, it's install the gcc-4.8 package and then update the 'gcc' alternative to make that one the default gcc 17:01
japhb tadzik: Ah, thanks. You mentioned checking it after a movie, and I was thinking "Wow, heck of a long movie!" ;-) 17:02
ajr_ DrForr - The process manager showed around 700 of the available 735MB at the highest point I saw. My attention was probably distracted by some exciting paint drying.
RabidGravy - I just updated to what I believe is the latest iteration of the OS, but I'll check for the gcc version. 17:03
RabidGravy I think the gcc version is crucial here 17:04
I'm just updating a windows machine that hasn't been updated for a year and a half 17:05
japhb tadzik: I am actually starting to use it in my Terminal-Print example: github.com/japhb/Terminal-Print/bl.../rpg-ui.p6 (That's not PR'ed back to the upstream yet, because I'm using development of the example to drive API and performance improvements to Terminal-Print itself.)
RabidGravy it's rather tedious
DrForr You're in better hands than mine, as I said I haven't tried yet. I got a new RPi from OSCON swag so I'll probably try later this week.
ajr_ RabidGravy: What are the symptoms of the gcc problem? What error messages at what stage? 17:08
RabidGravy I can't remember to be honest
it was like a year and a half ago :)
ajr_ Never mind, I was just hoping to whittle down the possibilities. I'm lucky to remember stuff a week and half ago. :-)* 17:09
stmuk_ ISTR the rPI gcc was quite dramatically broken
RabidGravy but with the gcc version above and an otherwise stock Raspbian it builds fine 17:09
stmuk_ I think it was either signal 11 or an internal gcc error 17:11
RabidGravy yeah I think it was some internal gcc error regarding some strange asm macro 17:12
RabidGravy actually yeah, it was a code generation error of some kind whereby it got confused while trying to make the asm stage 17:13
stmuk_ this windows 10 ubuntu subsystem thing is quite good 17:19
although rakudo doesn't build under it 17:20
DrForr They don't like competition :) 17:23
RabidGravy is it somewhat like coLinux was a few years back, with a Linux kernel running as a Windows service? 17:32
mst I believe it's more like the way e.g. netbsd's linux ABI emulation modules work 17:33
geekosaur understands it as running *alongside* the windows "kernel", since it's a microkernel arch 17:49
geekosaur which is why it gets to do things windows can't, like true fork() 17:51
qub hello! how to reset eof "flag" on $*IN? (in case of recieving multiple EOFs) - i tried binary "read", but it didn't help - pastebin.com/K1kUk4aq
geekosaur eof is not a character, you can't "eat" it 17:52
moritz can can reopen STDIN though 17:54
s/can/you/
mst geekosaur: oh, hmm, possibly 17:55
moritz used to work with $*IN = open('-') but that doesn't seem to work anymore; dunno how it's done today
geekosaur that may not b what they are seeing, I don't know. stdio often does remember eof on terminal because people get confused when eof isn't really eof
maybe libuv does too
ajr_ BTW, what's the justification for the "put" command? It's not obvious what makes it different from the other output commands. 17:59
DrForr The inverse of 'get'. 17:59
tadzik japhb; hah, yeah, now I remember! Sorry, I totally forgot then 18:01
mst tadzik: did you play with timotimo's POSTBUILD yet? 18:02
tadzik japhb: I was mostly concerned if it didn't accidentally break some existing functionality, since the module has no tests
mst: no, but the patch looks a lot like my AFTERPARTY from years ago :) Can't find that one anywhere though
qub moritz: tried to reopen pastebin.com/8vsgpdDh, but got error "read string requires an object with REPR MVMOSHandle" 18:05
geekosaur no, that would not work, stdin does not have a determinable path 18:06
qub but docs.perl6.org/type/IO$COLON$COLONSpecial 18:08
geekosaur the process inherits a file descriptor, it has no idea where that fd came from. and it may not *have* a path if it's e.g. a pipe or a socket
geekosaur also after $*IN.close there is not going to be a way to get it back 18:10
(ok, I can see IO::Special but I suspect you need $*IN.path.Str. and you *do* need to not .close, or nothing will save you) 18:11
"let me just go back in time and not have closed it, so I can reopen it even if it stopped existing like a pipe or socket does"
qub in case of .path.Str "open" tries to open file in working dir, but still no luck with open without close: pastebin.com/xC1dsMei 18:14
geekosaur if I understand IO::Special correctly that would seem buggy 18:21
what exactly are you trying to reopen / read past eof on, anyway? 18:23
qub next "file" 18:24
geekosaur ...
what next "file"
is this a socket, is this a pipe, is this a terminal, is this a magtape, is this a ??? 18:25
qub i have endless stream on stdin in which files are separated with eof char (ASCII 04) 18:26
geekosaur uh 18:27
geekosaur you still have nto answered my question. and ASCII 04 is not actually EOF. it is a command *to the terminal driver* that causes it to produce a short read to the program 18:27
Hotkeys m: say grep * < 1000, map * ** 2, (1, 3 ... *) 18:28
camelia rakudo-moar 9a6379: OUTPUT«(...)␤»
Hotkeys m: say eager grep * < 1000, map * ** 2, (1, 3 ... *)
why doesn't this work?
camelia rakudo-moar 9a6379: OUTPUT«(timeout)»
jast because it applies the test against an infinite series 18:29
that's my guess, anyway. grep doesn't know that it can abort after the first test that returns false...
Hotkeys fair 18:30
Is there a convenient way to do squares up to X with a sequence alone?
lizmat how many do you need?
Hotkeys (1 ** 2, 2 ** 2 ... $x) essentially 18:31
lizmat Hotkeys: slice it 18:32
m say (1,2,4...*)[^10]
m: say (1,2,4...*)[^10]
camelia rakudo-moar e42b68: OUTPUT«(1 2 4 8 16 32 64 128 256 512)␤»
lizmat m: my $x = 20; say (1,2,4...*)[^$x] 18:33
camelia rakudo-moar e42b68: OUTPUT«(1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288)␤»
jast that works if the number of elements is known; in this case it would happen to work because you can determine the correct limit using a logarithm 18:34
Hotkeys specifically the sequence I'm doing is squares of odd numbers actually so 18:35
jast ...[^log(1000)/log(2)]
Hotkeys 1², 3², 5² ... 1000
jast no wait, that doesn't make sense :) 18:35
sqrt(1000) actually
qub geekosaur: i need to enter arbitrary text line by line in terminal and send it to program by pressing ctrl+d 18:36
jast I'm too tired to think about this
qub: so just keep reading and whenever you get a short read, you know that the next read will start another file. unfortunately there will be no way to signal "this is the last file" but if your stream truly is infinite that shouldn't matter 18:38
Hotkeys does p6 have a function equivalent to takewhile? 18:40
timotimo Hotkeys: map with "last"
geekosaur jast, I think the point is that it stops after the first eof, which is a common thing that levels above the system call (such as libuv) do on terminals so that xxx^D works as expected by naïve programs) 18:41
but I suspect this is one of those things where we need to do something to libuv, and can't 18:42
jast I just tried with a tiny P5 program, I can keep reading beyond the first ^D
oh yeah, libuv, didn't think about the internals :)
qub yes, i tried to port p5 script 18:44
dalek osystem/MARTIMM-patch-1: 1d00beb | (Marcel Timmerman)++ | META.list:
Pod::Render module

Pod::Render module and pod-render program to render pod to html using Pod::To:HTML and markdown using Pod::To::Markdown. First can also be converted to pdf using wkhtmltopdf when installed. The module is a sort of wrapper to add better css support.		 18:46
FROGGS m: say map { .² }, (1, 3, 5 ... sqrt(1000)) # would be nice if *² worked too 18:47
camelia rakudo-moar e42b68: OUTPUT«(1 9 25 49 81 121 169 225 289 361 441 529 625 729 841 961)␤» 18:47
Hotkeys oh that is nice 18:49
pmurias hi 18:51
dalek osystem: 1d00beb | (Marcel Timmerman)++ | META.list:
Pod::Render module

Pod::Render module and pod-render program to render pod to html using Pod::To:HTML and markdown using Pod::To::Markdown. First can also be converted to pdf using wkhtmltopdf when installed. The module is a sort of wrapper to add better css support.		 18:57
osystem: a49baa0 | Altai-man++ | META.list:
Merge pull request #265 from perl6/MARTIMM-patch-1

Pod::Render module
japhb .tell tadzik I ran a few ad-hoc tests of both the new and old functionality, and fixed a couple bugs because of that ... did you want me to create a test file as part of the PR? 19:18
yoleaux japhb: I'll pass your message to tadzik.
tadzik japhb: nah, no need :) looking at the PR right now
yoleaux 19:18Z <japhb> tadzik: I ran a few ad-hoc tests of both the new and old functionality, and fixed a couple bugs because of that ... did you want me to create a test file as part of the PR?
japhb Oh, cool beans 19:19
tadzik hm, maybe a line for example.pl though? 19:22
japhb tadzik: Yeah, can do. 19:27
tadzik ossum
japhb tadzik: Looks like you merged the first one, so I made the example.pl extension a new PR 19:56
lolo78 Hi everyone. 21:29
A person that I know wants to rewrite his Perl 5 application in Perl 6. Part of it is a GUI written with Web technologies and using HTML5+ Canvas + Jquery, etc. I have no experience with Web development, especially not in P6. Is there any module or framework that you would recommend?
moritz lolo78: the state of P6 web frameworks is not very advanced right now; I guess my advice would be to use Inline::Perl5 (or Inline::Perl6 from p5) to gradually migrate business logic to p6, but keep using whatever p5 webframework the app i susing now 21:32
*using
zostay what does this mean? "Cannot invoke this object (REPR: Null; VMNull)" 21:33
moritz zostay: it means that you're trying to invoke something, but that somthing is a NULL on a very low level 21:34
geekosaur something internal leaked out into perl6 level 21:35
lolo78 Moritz: Thanks Moritz. Right now, I think the P5 application is using CGI, so it might be a good idea to upgrade to more modern technologies. If there is no framework, maybe there are still a few modules that can be used. It's not a full web application, but an application using a Web browser and HTTP for its graphical interface. Maybe a couple of good modules might be sufficient. I was... 21:37
...thinking about things like github.com/gfldex/http-server-simple or similar things.
zostay that's kind of what i figured, any suggestion on how i can track it down? 21:38
RabidGravy lolo78, I've been using HTTP::Server::Tiny + Crust for a few things, works well for me 21:39
moritz zostay: do you have a backtrace? if not, try with --ll-exception
zostay nope and nope then either 21:40
RabidGravy zostay, the pre-eminent cause of that is something like a .wrap or other low level operation on a callable that would cause it to get saved as an attribute during compilation
and the resulting closure "disappearing" during the pre-compilation
lizmat m: use nqp; (nqp::null)() # zostay 21:41
camelia rakudo-moar c5c600: OUTPUT«Cannot invoke this object (REPR: Null; VMNull)␤ in block <unit> at <tmp> line 1␤␤»
RabidGravy there are tickets for it in RT
zostay RabidGravy++ i am doing something like that and commenting out bits of that code at least change what's happening in a way that might help me find, thx 21:43
qub regarding non-working multiple eof-s in $*IN: where should i send the bug report: rakudo/nqp/moarvm/libuv? 21:44
moritz qub: rakudo
zostay Perl 6 is such a joy to play with when it's a joy. When it isn't, it's like being nibbled to death by cats.
moritz that's where bug reports go unless you're quite sure what layer it's in 21:45
qub: that said, what exactly is the desired behavior?
RabidGravy regarding that, I think it was BMC Remedy that stored custom comment fields in a large BLOB in the database separated by FF characters 21:48
RabidGravy there's no end of crack addled behaviour in software that originated in the pre-modern era 21:49
I've just done a test on Linux and it will happily "$*IN.slurp-rest' something with an embedded 004 in it 21:53
geekosaur yes, because that is not actually EOF
geekosaur this is a tty driver-specific behavior, the stty-d eof char causes a short read 21:53
the tty eof char is not passed on 21:54
RabidGravy right, but then I'm not sure what we are talking about
geekosaur stdio, and probably libuv, make eof "sticky" on terminal filehandles, because the common idiom of looking only for an 0-length read will fail if you type xxx^Dyyy
geekosaur (that is, programs will handle the initial short read as a normal read that returned less than usual, and expect the next read to return 0. with ttys the next read returns more data, so libs like stdio and probably libuv remember they saw the short read and return EOF from then on unless "reset" 21:55
RabidGravy right, so how is it in the data that qub is talking about then? 21:56
geekosaur qub's complaint is they are sending data via terminal, control-d, then the next set of data, control-d, ... 21:56
the short-read workaround prevents this because after the first short read, the handle is considered to always be at EOF 21:57
RabidGravy well, it's just stick the terminal in "raw" mode with stty then
geekosaur but then you can't detect end at all unless you watch for some magic sequence at the beginning. also you will need some way to pass through a control-D that is *not* an eof
basically, replacing one can of worms with a smellier one
lizmat perhaps an IO::Handle.split( FF ) could work ? 21:59
RabidGravy I'm feeling the hate growing, despite having maintained a large proportion of the P5 Term::* modules into the 21st centurt
y
geekosaur thinks the thing being done here could be done better some way...
tar|uuencode and uudecode/split on the remote, or something 22:00
RabidGravy or, y'know providing a sane modern interface or something 22:01
VVbrJVldCqK www.youtube.com/watch?v=3EsJLNGVJ7E & wikileaks.org/podesta-emails/emailid/15893, www.reuters.com/article/us-usa-elec...SKBN12Z2SL & wikileaks.org/podesta-emails/emailid/3774 (ctrl+f qatar) - please don't let these be buried 22:04
RabidGravy wikileaks being an agent of the russian state now and all 22:05
twats
right, that's me done for the day, have fun 22:06
lizmat good night, RabidGravy
lizmat wonders whether we should auto-purge any drive-by messages like the one from VVbrJVldCqK just now 22:13
basically the ones with /join /msg / quit and nothing inbetween 22:14
(from the clogs, I mean) 22:15
TEttinger but that would bury the spam! and it asked not to bury the spam with a "please" 22:19
grondilu if the message also contains only URLs, then yes. 22:45
oh there was an additional request not to burry the thing indeed 22:46
(my bad)

Powered by Cro and the Raku® Programming Language.

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

Thank you!