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.
mncharity [particle]: a yaml dump of languages/perl6/perl6 --target=past would be great. Curious that --target=past ignores -e. 00:33
pmichaud (--target=past ignores -e) I noticed that also yesterday. Probably deserves a ticket to [email@hidden.address] 00:34
Juerd pmichaud: Hi. I see you're a speaker at fosdem tomor... eh, in little over 12 hours :) 00:38
pmichaud: Do you know if there's any Perl community gettogether?
use.perl.org/comments.pl?sid=38508&cid=60912 # acme suggested meeting too, but without location and time 00:46
lambdabot Title: FOSDEM
pmichaud I'm not aware of a perl community get together... I probably should've tried to organize one. :-| 00:50
I'm pretty busy saturday (tomorrow), but so far sunday seems open 00:51
mncharity re 'Probably deserves a ticket to [email@hidden.address] sent.
pmichaud mncharity++ # thanks
Juerd: I've asked others, but can ask you also -- any particular points to be sure to hit/avoid about Perl 6 in tomorrow's talk? 00:52
Juerd pmichaud: That people shouldn't compare software based on version numbers. 00:54
(hit)
Especially since Perl 6 and PHP 6 are scheduled so competatively 00:55
PHP 6 will probably be released long before Perl 6, but Perl 6 is an almost new language.
pmichaud really? When is PHP 6 coming out? 00:56
Juerd Probably before it's finished ;) 00:58
But probably also before christmas
Didn't you know that the PHP 6 talk was scheduled directly after yours? 01:00
In the same room even
pmichaud yes, I did know that 01:01
pmichaud but I hadn't linked "Perl 6" and "PHP 6" in my mind. You're correct that many audience members might. 01:01
(I had already planned to attend the PHP 6 talk, since I have a few PHP projects) 01:02
Juerd Ah. It is my experience that many people think version numbers of certain kinds of software are almost the same because it's some kind of competative race.
Juerd Internet Explorer and Netscape, versions 3..5 had the same problem. 6 less so. 01:03
Some say that PHP released 5, which isn't really that big a difference since 4, to get in line with Perl. It's probably false, but still there's people who think this.
pmichaud well, I know that I did see similar things with Linux releases for a while (e.g., between RedHat and SUSE) 01:04
Juerd The PHP 6 talk is "Unicoding with PHP 6" by the way.
But on the printed schedule it's just titled "PHP 6"
pmichaud so yes, I know what you're talking about. I'll see if I can gently make the disconnection in my talk. But my talk already focuses on Perl 6 being an altogether new language in many ways 01:05
Juerd Ah, that's great in any case :)
'cause That appeals to both Perl lovers and Perl haters :)
Juerd Many Perl users like learning a new powerful language, especially if it is "Perlish" 01:06
pmichaud a little over half of the talk is "what is Perl 6, why it is great", the remaining is "how are we implementing it (in Parrot)"
Juerd And Perl-haters will see that many of their complaints, except noise about line noise, is gone :)
pmichaud yes, I already cover the sigil changes 01:07
Juerd I think that that is a good format for this audience
pmichaud I gave a similar talk to this one a few months ago, and it ran 90 minutes. So I have to figure out how to get down to under 45. :-)
but I'll do it. :-)
Juerd Are you going to mention just Rakudo, or the multiple implementations?
pmichaud definitely mention the multiple implementations
Juerd Or is this too sensitive a subject to mention on IRC :D
pmichaud I always mention that there are multiple Perl 6 efforts 01:08
Juerd got a few off-list responses to his "focus on a single one for now" suggestion
pmichaud the rakudo implementation will be interesting because it's using the parrot compiler toolkit (which is a significant talk in its own right) and because it really is written using Perl 6
Juerd I forgot to mention that that statement was based on dollar based development. 01:09
pmichaud yes, but in many days dollars and effort are interchangable
personally, I'm very glad we have multiple efforts taking place. I also don't see it as zero-sum
Juerd I am also glad we have multiple simultaneous approaches 01:10
However, for funding "Perl 6" I'd say it's best to pick one at a time.
pmichaud my experience is that the multiple efforts are making it easier for all to progress. I think that's actually necessary in a project as ambitious as Perl 6
Juerd It appears not many people share that view.
pmichaud well, much of the funding that has been available thus far has been targeted 01:11
(to parrot). That's not a function of TPF or the grants committee, but rather the organizations contributing the funding
SamB well it's definately better then having all the effort spent in barking up the wrong tree, and possibly better than leaving some people who'd otherwise contribute out since they can't find anything to work on...
Juerd The more transparent the ambition (as in "we've never done it before, and don't really know if it'll work the way we want it to") is to investors, the fewer bucks they'll spend, I think. It's a gut feeling.
pmichaud if other organizations want to contribute to other efforts, that's perfectly okay also.
SamB doesn't think he'd ever consider giving money to something with that many opcodes... 01:12
pmichaud and by "other organizations" I mean other organizations that want to contribute to other implementations via TPF
Juerd pmichaud: That effectively forces sponsors to dive into Perl 6 culture, which raises the bar a lot.
pmichaud Juerd: right. The organizations that have made substantial donations, such as NLNET and Mozilla, dove into the culture enough to see that the Parrot effort was where they wanted to focus their efforts 01:13
I don't know what is happening to the "non-targeted" funds that TPF grants may have available. 01:14
Juerd pmichaud: Those organizations have their roots in open development 01:14
It's easier for them.
pmichaud correct
if people look at those grants and say "TPF is only supporting Parrot", that's a mistake. 01:15
TPF is making it possible for those other organizations to support Parrot.
or at least making it easier.
Juerd I know 01:16
And I'm not objecting
I'm looking at it in the context of Conrad's thread starting mail 01:17
Which was about raising funds for Perl 6, not about TPF or past grants.
And I do think that it's easier to ask corporations for their money to fund a product, rather than something resembling a research project. 01:18
pmichaud yes, it is.
Juerd Apart from that, I agree wholeheartedly that having multiple implementations is good, and that TPF should be able to fund any and all of them.
pmichaud right -- that's really up to the grants committee.
Juerd I had no idea that there was still 35k left 01:19
avar [particle]: re svn, no, a .git:) I keep it at git.nix.is
[particle]: git-clone git://git.nix.is/avar/pm/re-engine-TRE
Juerd And to be honest I'm a bit surprised, since I've seen people complain "We're unpaid volunteers, so duh, it'll take a while" while the "unpaid" part turns out to be by choice... 01:20
I'm not suggesting that everybody should start taking money, but I am under the impression that several people who would like to spend more time on hacking on Perl 6, had no idea that their work had a chance of getting funded. 01:21
pmichaud well, the NLNet funds aren't totally unencumbered -- they're restricted to specific milestones
Juerd (Which would not make being unpaid a choice - it would make it a result of ignorance)
pmichaud so we can't just say "here's $x to go work on whatever you think needs working on"
Juerd Ah.
Unrenegotiable?
pmichaud it was renegotiated once... but there are some broad limits to what the funds are intended to cover, yes 01:22
Juerd No, not "whatever you think", but wouldn't it be useful if people could at least suggest alternative milestones?
Of course, bureaucracy might kill the process. I have no idea how to fix that.
Anyway, I should go home and get some sleep before I drive to Brussels tomorr... today. 01:23
pmichaud I should finish my presentation slides. :-)
Juerd Good luck :) 01:25
pmichaud thanks. :-)
at least I have a head start on this one. Many times I only have 3-4 hours to finish.
right now I have about 7 or 8. 01:26
(and a head start on the slides from my previous presentation, also :-)
meppl good night 01:34
mncharity [particle]: spinclad: ok, yes, ../../parrot perl6.pbc --target=past should definitely be included as a parser source. especially once it gets a yaml output form - I'm unclear on whether there is p5 or other code to pick up the existing output. 01:50
TimToady mncharity: how'd the meet with obra go? 01:53
pugs_svnbot r19982 | lwall++ | remove :: from filenames for windersboxen 02:06
diff: dev.pugscode.org/changeset/19982
lambdabot Title: Changeset 19982 - Pugs - Trac
csharp download links don't work ~ www.pugscode.org/ 02:23
lambdabot Title: Pugs - pugscode
csharp is it just me? 02:25
pugs_svnbot r19983 | lwall++ | [gimme5] default to not failing over if lexer fails (for now) 02:28
diff: dev.pugscode.org/changeset/19983
lambdabot Title: Changeset 19983 - Pugs - Trac
TimToady doesn't work, just pull it from svn to get the bleeding edge (which requires GHC 6.8.1 I think at the moment) 02:32
csharp yeah, i just found the page for svn co (www.perlfoundation.org/perl6/index....load_pugs)
lambdabot Title: Download Pugs / Perl 6
mncharity TimToady: re meet, useful. () he thought getting us towards a documented stable ast which people could target was the main potential contribution. () i found it interesting how several things i've come to think of as obvious, even in extended conversation with someone familiar with the area, proved not so. () got me over threshold to revisit parrot. 02:36
() it was interesting the difficulty of characterizing what i've come to view as the critical importance of a fast edit-test cycle. given a _work_ p6 snatched from the future, however slow, we would be all set. but given existing buggy and incomplete implementations, slow is deadly. 02:38
* _working_ 02:39
there were some others... hmm... 02:40
() oh, very interesting hearing of views from the broader p6, and p5 viewing p6, communities. which i almost never see. 02:42
mncharity () he found... non-obvious?... my perspective that the p6 critical path has been on hold for several years (modulo various projects contributing to language spec evolution), since pugs's first summer, blocked by not having a parser willing to provide a parse tree which includes oo info. thus killing backend and prelude development, and thus any large-scale use of p6. 02:48
mncharity highlights. 02:53
looking at parrot's language/perl6, some of the choices are interesting. eg, there's support for junctions and (reportedly - blog) for roles, but not for Array's. 02:58
s/choices/emphasis/. "bottom up"-ish. 03:00
perhaps.
TimToady back from dinner 03:05
thanks
pugs_svnbot r19984 | lwall++ | change bad download link in at least one spot... 03:16
diff: dev.pugscode.org/changeset/19984
lambdabot Title: Changeset 19984 - Pugs - Trac
pugs_svnbot r19985 | putter++ | [kp6] added -C ast-yaml option. 03:22
diff: dev.pugscode.org/changeset/19985
lambdabot Title: Changeset 19985 - Pugs - Trac
pugs_svnbot r19986 | putter++ | [kp6] Changed -C ast-yaml from YAML::Syck to YAML::XS. 04:49
r19987 | lwall++ | [gimme5] much cleanup of trace output
diff: dev.pugscode.org/changeset/19987
lambdabot Title: Changeset 19987 - Pugs - Trac
TimToady mncharity: what is the difference between YAML::Syck and YAML::XS? 04:57
mncharity [particle]: apropos kwalify, regrettably it doesn't deal with tags (nor permit enum's of maps_. so, it doesn't seem to help us. so... I'm thinking of metaprogramming something which loads the yaml and grovels over the resulting objects. :/ seems less painful than going the xml route. 05:01
TimToady: re Syck vs XS... just a past end-of-day mistake. jumping at unicode bug shaddows. reverting. 05:13
pugs_svnbot r19988 | putter++ | [kp6] revert r19986. Syck seems vaguely preferable to libyaml/XS on both perl and ruby. But it's late, so this may get revisited, again. 05:22
diff: dev.pugscode.org/changeset/19988
lambdabot Title: Changeset 19988 - Pugs - Trac
mncharity /me nodding off. good night all & 05:23
TimToady @tell mncharity yes, it would be very nice if any fast form of yaml had its unicode act together... 05:28
lambdabot Consider it noted.
pugs_svnbot r19989 | lwall++ | [gimme5] getting --> coercions to add in prec and assoc to %thisop 07:14
diff: dev.pugscode.org/changeset/19989
lambdabot Title: Changeset 19989 - Pugs - Trac
Tene pleasedieinafire.net/~tene/grammar.png -- 20% 09:40
pmichaud Tene: that's for rakudo? 09:42
Tene Yeah.
20% of rakudo's grammar.pm 09:43
pmichaud ah
<afterws> isn't called? or just isn't linked yet?
Tene nothing up to that point calls it.
pmichaud actually, an interesting grammar to graph might be NQP's grammar :-)
it's a bit smaller 09:44
still, this is quite cool
pugs_svnbot r19990 | lwall++ | [gimme5] EXPR more stable, much debugging of endsym logic 09:46
diff: dev.pugscode.org/changeset/19990
lambdabot Title: Changeset 19990 - Pugs - Trac
TimToady what's a routine_declatorot? :) 09:47
Tene A typo, of course.
pmichaud it's the official name for bit decay
bit decay that occurs regularly, that is
Tene 33% done 09:49
Tene There, done. 10:06
the source is pleasedieinafire.net/~tene/grammar.src
compile with: dot grammar.src -Tpng -o grammar.png
pmichaud unfortunately, <TOP> isn't at the top. 10:09
Tene Heh. 10:10
Hmm, wonder if i can fix that.
pmichaud internally, optable calls <term>
Tene re-rendered with optable -> term 10:13
And that put TOP at the top.
pmichaud 'quote_expression' calls 'quote_concat' 10:15
'quote_concat' calls 'quote_term'
'quote_term' calls 'variable' 10:16
and 'quote_literal'
(those are all in src/parser/quote_expression.pir, written as PIR instead of regexes 10:17
Tene updated
Now TOP is in the middle of the top. Entertaining.
pugs_svnbot r19991 | lwall++ | [gimme5] getting endsym info up to where autolexer can see it 11:05
diff: dev.pugscode.org/changeset/19991
lambdabot Title: Changeset 19991 - Pugs - Trac
rakudo_svn r26016 | paultcochrane++ | [rakudo] Updating svn properties appropriately so the file_metadata test runs 11:52
r26016 | paultcochrane++ | correctly for svk working copies.
Juerd pmichaud: Great talk :) 14:06
mncharity does Moose provide multi-methods? I've been groveling over the docs and do see them. Is there anything now available aside from Damian's very old Class-Multimethods, and Lukes "one global namespace" Class-Multimethods-Pure? 16:34
*don't
SamB considers making a stupid joke about MOOSE being a convenient way to edit MOO objects 16:52
mncharity Moo Software Engineering? 16:58
TimToady I don't believe anyone has actually implemented the currently specced multi dispatch 17:09
because MO is necessary but not sufficient
multiple dispatch tends to live outside the purview of any one type 17:11
though of course the Code types have to deal with it on various levels 17:12
TimToady it's another one of those areas that everyone else has been avoiding thinking about, like longest-token matchers :) 17:13
so I wouldn't be surprised if that was something I had to prototype after STD is "done" 17:14
but by all means, feel free to work on it! :D 17:15
mncharity :P /me dodges the concept by clarifying that it was just a normal "I want to do X in p5. The RightThing approach is multiple dispatch. Before I decide on which not-RightThing to use instead... time has passed... I don't suppose there's a chance that...". 17:21
though ended up with a not-RightThing of single dispatch on typename suffix, which has the nice property that
it permits ducktyped dispatch on ast nodes from different sources... ah bleep... it's all kludge. 17:23
mncharity [particle]: re rakudo ast yaml, any idea when it might be available? 18:13
pugs_svnbot r19992 | lwall++ | [STD] rewrite panic as normal method because it's hard to <commit> 20:32
r19992 | lwall++ | [gimme5] properly recognize (and ignore <rw> on context)
r19992 | lwall++ | [Cursor5] double-quoted string now pulls out fixed prefix, if any
r19992 | lwall++ | [Cursor5] only non-atoms in quantification context need extra grouping
diff: dev.pugscode.org/changeset/19992
lambdabot Title: Changeset 19992 - Pugs - Trac
tewk 21:38
pugs_svnbot r19993 | lwall++ | [gimme5] various metachar and yaml vs. unicode issues 22:02
diff: dev.pugscode.org/changeset/19993
lambdabot Title: Changeset 19993 - Pugs - Trac 22:03

Powered by Cro and the Raku® Programming Language.

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

Thank you!