pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
pugs_svnbot r20174 | lwall++ | [STD] removed whatever token which duped term:sym<*> 01:17
r20174 | lwall++ | various warning suppressions and lexer simplifications
diff: dev.pugscode.org/changeset/20174
lambdabot Title: Changeset 20174 - Pugs - Trac 01:18
meppl good night 02:53
syle request, can we keep c type syntax, perl6 is looking like ruby and other crap 05:22
syle its never been hard to code perl having c/c++ background , now code is looking like some tcl language 05:24
Khisanth that doesn't sound like a very compelling reason 06:06
spinclad oh, hardly like tcl. believe me, i know from tcl. 06:15
(though Coke, over on #parrot, knows far better than me.) 06:16
syle what are thoughts on parallel computing? 06:30
pmurias mncharity: hi 11:21
pmurias there seems to be a bug or a at least strangeness with the Match objects dumped by STD_red_run 11:25
the str field referes to the part of the string left instead of the orginal string 11:26
implemented the string sharing code but i didn't commit it as it makes no sense with the current design 11:27
cognominal_ TimToady: %a = { key => @a } # is a array (or hash) as a value is allowed in a hash composer? I suppose that in that position we don't expand it. 13:37
ruoso cognominal_, I think it's a matter of precedence, and who enforces the context... if it's the pair who enforces item precedence on the array, or the array which is expanded before that... 13:48
I would argue that => is of a higher precedence, and enforces item context on the array
s/item precedence/item context/ 13:49
alester Who is Mitchell? 14:09
www.rakudo.org/2008/03/perl-6-desig...9-mar.html
lambdabot Title: Perl 6 Design Notes for 19 March 2008 - Rakudo.org
pmurias mncharity 14:13
Mitchell N Charity, known in commit messages as putter 14:14
alester Never heard of him. 14:19
spinclad he's working on 'elf' (another bootstrap project, p6 to p5/ruby, STD parser to a commonish AST) and project management of this herd of cats 14:27
hang around here and you'll meet soon enough 14:28
alester ok 14:59
mncharity pmurias: hi 17:36
the str field refers to the part of the string matched by the Match, no? 17:37
the alternative is, at least in yaml land, to have an on_str field, which points to the original string, with yaml taking care of making that a reference rather than a copy, and Match on the p5 side then recreating str, either when created or on demand. 17:38
mncharity for --dump5, the simplest thing might be to just do my $on_str = ...; match(...$on_str...etc...); 17:40
then Match under elf_a and elf_b would have to be set up to derive str from on_str plu from and to. 17:41
mncharity I may have just missed it, but what are we buying? ie, what is the need (reducing --dump5 size? is it a problem?); how much does the solution address it? (ie, what is the payoff - eg, 40x size of source -> 20x?). 17:49
mncharity saw a nifty talk yesterday on programming using a continuous spacial fields abstraction over meshes (sensors, robots, articulators, whatever). no link. :/ author of goo (language). so much to do... 17:55
pmurias mncharity: already coded that for --dump5 18:06
i think the .str should refere to the whole string matched on 18:07
* refer
re coded, the dump not the reading in 18:08
pmurias it's possible to extract the part the string matched on using .from and .to 18:11
TimToady j 20:00
[particle] l
PerlJam i 20:13
b_jonas k 20:16
kolibrie h 20:17
diakopter heh 20:57
mncharity lol 21:35
pugs_svnbot r20175 | putter++ | [elf] elf_b can now create and describe IR trees.
r20175 | putter++ | [STD_red] bugfix r20172 sigil escaping.
diff: dev.pugscode.org/changeset/20175
lambdabot Title: Changeset 20175 - Pugs - Trac 21:35
mncharity elf_b next step is ir -> emit_p5. when that works, we'll technically have a backend bootstrap - a compiler demonstrably capable of compiling its own backend (ie, everything but STD_red). 21:41
mncharity elf_b is tightly tied to elf_a. so the step after that may be to use elf_a/b (while continuing to tweak them), to create elf_c, also a bootstrapped backend, but standalone and free of the "derived from elf_a" constraints and scary magic. 21:44
mncharity longterm question: does anyone know of a language which gives useful line numbers despite having macros? one basically wants the whole macro expansion history, for every character position, available when describing errors. 21:51
b_jonas mncharity: uh, the full macro expansion history? interesting 21:52
that would mean a quadratic stack trace 21:53
like
died: foo problem (FooException)
from functoin1 prog.ext line 1 char 2 (expanded from line 3 char 4, line 5 char 6, ...)
called from function2 ... 21:54
that would be huge
mncharity yes
b_jonas no, I don't know of anything like that
also, ask for implementation instead of language 21:55
because e.g. scheme implementations differ very much in how good error messages they give
(though they differ greatly in other things too, but still)
mncharity but otherwise if foo($x.bar) {} , where statement_control:if and foo are macros, easily results in macro expanded p6 code, and then in compiled runtime code, where the connection between some error and that line is... obscure. 21:56
re implementation, good point
b_jonas well, you could easily supress "system" libraries 21:57
even if that would actually mean gaps in the expansion line
that's not confusing for you can think macros in system libraries are primitives
obviously, they would be supressed only unless you set a certain compiler option 21:58
perhaps ask on #scheme or #lisp 21:59
because those are the languages that have macros
apart from the C preprocessor
mncharity the p6 prelude is rather unprivileged. so what you are describing is the ability to elide info from particular modules. which is a good idea, but doesn't make the issue go away... 22:00
b_jonas yep
by "system", I meant like the "system" headers in C
so whatever claims itself to be system 22:01
and bug-free
mncharity re scheme/lisp, yeah, and check on what dylan ended up doing. 2nd(?) meeting of the Boston Lisp Users Group on Monday. :)
b_jonas what's the connection between dyland and boston list users group?
s/dyland/dylan/
which is dylan btw? is that that other language with builtin syntax for coroutine-based iterators? 22:02
mncharity scheme/lisp ~~ dylan, and BLUG is an opportunity to ask a bunch of scheme/CL/dylan folks, including several implementers, for thoughts. 22:03
re dylan?,
en.wikipedia.org/wiki/Dylan_programming_language
lambdabot Title: Dylan (programming language) - Wikipedia, the free encyclopedia
b_jonas that's not the one I was thinking than 22:04
what language was that then? 22:05
wait, it might be the same
ah, I remember this
I started to read the reference manual but stopped soon because it's so dry as a standard 22:06
even though I like reference manuals
b_jonas I can't follow all these languages 22:08
but yes, dylan has iterator syntax too 22:09
just like python and ruby 22:10
smalltalk++ for popularizing iterators
mncharity dylan was sort of an attempt to be CL/scheme-next-generation, and be java (in the 'next big-time computing language') sense of 'be java'. so it has a lot of carefully thought about and documented ideas. so it can be a useful thing to check when considering how to do X where X is something lispish. 22:13
b_jonas I see 22:23
pugs_svnbot r20176 | putter++ | [elf_b] Copy elf_a's emit_p5.config, for hand translation into p6. 22:54
diff: dev.pugscode.org/changeset/20176
lambdabot Title: Changeset 20176 - Pugs - Trac
mncharity anyone know what the current terse form of self.foo() is? at one point it was ./foo() , but I'm not seeing that in the Snn 23:22
TimToady you're looking for $.foo() 23:39
or @.foo(), or %.foo()
shorthand for %(self.foo) etc. 23:40
the outer sigil doesn't have to match the attribute declaration, just be consistent with it 23:41
(or the method declaration)
$.foo() has the benefit of interpolating, unlike ./foo() 23:42
TimToady diakopter: thanks 23:43
mncharity ah, ok. thanks 23:49
diakopter Infinoid got a new svnbot running in #parrot 23:51
mncharity it even parses (in STD_red), albeit as a subcall.
extra features? 23:52