svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by lumi on 8 November 2006.
nothingmuch bah 00:00
nevermind
sleep & 00:02
00:05 chromatic joined
chromatic Assuming I have an hour and want to experiment with writing a Parrot emitter for MiniPerl6, where should I look? 00:06
Limbic_Region salutations chromatic
?seen audreyt
lambdabot audreyt is in #perl6, #ghc and #haskell. I last heard audreyt speak 3h 34m 18s ago.
Limbic_Region ?seen cmarcelo
lambdabot I saw cmarcelo leaving #perl6 6h 38m 35s ago, and .
nothingmuch *moose noises* 00:07
Limbic_Region chromatic - MP6 is relatively new so I don't know much about it and the two who know it best aren't around
nothingmuch - thought you were sleeping
chromatic I made a Parrot.pm with an empty emit {} in perl5/Pugs-Compiler-Perl6/, then invoked it with perl -Ilib -Mblib lib/v6.pm --compile-only -BParrot say42, but no luck.
nothingmuch moose noises in mysleep ;-)
chromatic say42 is a file with "say 42" in it.
nothingmuch i'm just post-hygene obsessing
closing windows etc takes long 00:08
plus that last commit is always a bit longer than expected ;-)
00:08 mako132_ joined
Limbic_Region though I might even be confusing MP6 with OM 00:09
nothingmuch last link: home.pipeline.com/~hbaker1/CheneyMTA.html
lambdabot Title: Cheney on the M.T.A., tinyurl.com/purol
nothingmuch very interesting for those who appreciate the low level side of high level languages 00:10
Limbic_Region chromatic I assume you have read the README in pugs/v6 ? 00:12
[particle] c: look in v6/v6-MiniPerl6/Emitter/
chromatic Yes. 00:13
Limbic_Region goes back to sleep then
chromatic lib/MiniPerl6/Emitter, you mean?
[particle] er, yes 00:14
chromatic In theory, then, I could copy this and override all of the emit* methods to emit Parrot code then? 00:15
[particle] yes
chromatic How would I test it?
[particle] i'm not quite sure 00:16
chromatic How can I tell it to use a different emitter? 00:17
[particle] i'd guess you'd want to spit out the pir at first, and sanity check it
chromatic Does the code even work? 00:18
[particle] yes, it works with perl5 back end 00:19
chromatic Okay. How do I make this code work with the P5 backend? 00:20
[particle] oh, i think maybe the emitter refactor isn't quite done yet
fglock got a new system, or has some system restore hassles preventing that commit
if you have v6 installed, just switch to v6-MiniPerl6 and run 'prove t' 00:21
chromatic t/01-sanity....FAILED before any test output arrived 00:22
t/02-emit......FAILED before any test output arrived
t/03-token.....FAILED before any test output arrived
t/04-regex.....FAILED before any test output arrived
FAILED--4 test scripts could be run, alas--no output ever seen
TimToady I don't think those tests are doing TAP format yet.
[particle] oh. well then....
chromatic $ perl -Ilib t/01-sanity.t 00:23
Error in perl 6 parser: invalid rule syntax at (eval 79) line 349.
$ rm t/*tc && perl -Ilib t/01-sanity.t 00:24
Error in perl 6 parser: invalid rule syntax at (eval 79) line 349.
$ rm t/*tc && perl -Ilib -I../../perl5/Pugs-Compiler-Perl6/blib/lib t/01-sanity.t 00:25
Error in perl 6 parser: invalid rule syntax at (eval 79) line 349.
[particle] ../../misc/pX/Common/Pugs-Compiler-Perl6/lib? 00:26
that's the bleeding-edge stuff
chromatic v6 from CPAN wouldn't install either. 00:27
I figured the new stuff might have a chance of working.
~/dev/pugs/perl5/Pugs-Compiler-Perl6 $ make test
1 test and 2 subtests skipped.
Failed 2/35 test scripts. 314/721 subtests failed.
Right then. Yet another half-assed, overpromised Pugs code-tease. 00:29
TimToady be nice
chromatic Larry, every time I try to do something on Pugs, it fails to work. 00:30
[particle] heh, and i thought it was just me 00:31
chromatic Every six months, someone says "Oh, try it now! It's so much better, and we could use your help!" and it fails to work.
[particle] i keep thinking there's a magic document on using this stuff, and i'm not reading it. but, it doesn't exist, and i can't use it.
chromatic After three or four or five times, I start to believe that it just doesn't work. 00:32
Limbic_Region chromatic - you aren't interested in the Haskell backend then? Just parrot and perhaps perl 5? 00:33
chromatic The code is impenetrable.
I gave up on the Haskell backend in February.
cj hey there chromatic
chromatic The Perl 5 backend has no documentation that I can find (or comprehend) and fails to pass the tests in tree or from the CPAN.
I decided to spend an hour experimenting today to see if I could make any progress with the Parrot backend. 00:34
Hi cj.
Ahh, that CJ.
cj ha!
Limbic_Region well, I don't hack internals - I just write perl 6 code. And when I tell pugs to run it - it does. A lot of it too, not just variable assignment. 00:35
00:35 nwc10_ joined
zgh yeah I've had good luck with actual perl6 code in pugs 00:35
HTTP::Server::Simple and CGI work together now ^_^ 00:36
yay web hacking in perl6!
TimToady I'm doing production Perl 6 code at work...
chromatic Yes, I'm sure the implemented stuff is great.
Limbic_Region there is a LOT of implemented stuff
most of t/ passes with the haskell backend 00:37
chromatic Right, it's slightly more capable than it was when I stopped working on it in August 2005.
When was the last time you ran make test? 00:38
Did it run in under two hours for you?
TimToady 40 minutes
Limbic_Region I run make test twice a day
chromatic That's positive.
Limbic_Region s/twice a day/at least twice a day/
on this box - < 40 minutes 00:39
on my slow machine at work circa 65 minutes
chromatic Is GHC 6.6 available for your platform?
Limbic_Region in neither place have I taken advantage of multi-processors
winbugs - yep
nor have I taking advantage of the ghc 6.6 performance flags 00:40
TimToady and that's with a lot more tests than it had a year ago.
chromatic I am leaving Flavio a note. I will only try once more. 00:43
00:43 bsb joined
Limbic_Region ?seen fglock 00:43
lambdabot I saw fglock leaving #perl6 4h 45m 56s ago, and .
00:46 chromatic left 00:57 weinig is now known as weinig|fooding 01:10 lyokato joined
follil ?seen audreyt 01:11
lambdabot audreyt is in #perl6, #ghc and #haskell. I last heard audreyt speak 4h 39m 23s ago.
01:21 dmq joined 01:22 pjcj joined 01:32 bonesss joined 01:37 Steve_p joined, avarab joined 01:46 fayland joined 02:04 fayland_ joined 02:09 nipotaway is now known as nipotan 02:10 bonsai joined 02:22 duend joined 02:45 qmole joined 02:51 mako132_ joined 02:57 Whoosh joined 03:05 weinig|fooding is now known as weinig 03:11 Coleoid joined
audreyt goes back from Rio.pm social meeting 03:19
oy. I missed chromatic. another chance of getting him to join development missed 03:20
c'est la vie.
pasteling "Coleoid" at 70.239.92.24 pasted "Changes to Makefile.PL to get it closer to working with Win32 $ENV{SVN_ASP_DOT_NET_HACK}" (16 lines, 664B) at sial.org/pbot/21001 03:25
Coleoid audreyt: This is the half-solution I mentioned earlier... 03:26
With that change, the generated Makefile needs only one more manual deletion to succeed (with groaning)... 03:27
Though more fixes are worthwhile.
GeJ adreyt: hi 03:28
Coleoid I don't know where the offending portion is being generated.
GeJ audreyt: having chromatic join the pugs ship would be great as he is undoubtedly a most valuable asset for perl 6 03:29
Reading the comments though makes me feel sad, though. I think I can understand his frustration, but his reaction is not helping. 03:31
Coleoid GeJ, I agree. I also empathize with his experience: I've tried to get up and working in Pugs several times.
GeJ Well, this is just the humble opinion of someone who has never provided anything to either pugs or parrot and as such can be discarded. 03:32
stevan GeJ: this is also just how chromatic
is 03:33
Coleoid This is the best I've done yet! With manual makefile hammering, I get a working build.
audreyt ooh! 03:34
"with" manual hammering?
Coleoid: also, please commit the Makefile.PL patch
Coleoid Yeah, after Makefile.PL is run, I've got to delete one more chunk. 03:35
(from the generated Makefile)
GeJ stevan: I never had the opportunity to meet him (and probably never will). All I know from him is what one can read on various blogs, MLs, and such. All I can say is that I have a high respect for him and his works. But the animosity shown in the comments he made makes me feel mostly uncomfortable. 03:36
Coleoid So the job isn't done yet, but I've never been inside MakeMaker. 03:37
(or any of the related techs.)
audreyt Coleoid: which chunk?
Coleoid Er...
pasteling "Coleoid" at 70.239.92.24 pasted "Manual hack to generated Makefile to get build success when working with Win32 $ENV{SVN_ASP_DOT_NET_HACK}" (10 lines, 418B) at sial.org/pbot/21002 03:39
Coleoid That chunk. 03:40
audreyt so it's trying to install _svn as a script
Coleoid: you got a commit bit, btw? 03:41
stevan GeJ: I wouldn't take the comments too heart too much, he means what he says, but I never found him to be mean spirited, just very spirited :)
Coleoid Mm-hm. There are other places this comes up in the Makefile, but that's the only fatal one.
audreyt it's now fixed.
GeJ stevan: good to know :)
Coleoid I don't have a commit bit.
svnbot6 r14684 | audreyt++ | * Makefile.PL: Coleoid observed that script/_svn/ was erroneously
r14684 | audreyt++ | being included in install_script; this fixed that.
audreyt oh. we can fix that. wha'ts your email address? 03:42
Coleoid [email@hidden.address]
clkao audreyt: now all you need is a restian api so you can give out commitbits without leaving your terminal..
luckily this web framework seemed to provide one..
audreyt Coleoid: welcome aboard; please add yourself to AUTHORS :) 03:43
Coleoid :D Thanks! 03:44
audreyt Coleoid: so please commit the Makefile.PL 03:49
and then it should build without fatal errors :)
Coleoid Merged in my changes, and I'm re-making to be sure... 03:50
audreyt Coleoid: sorry I can't stay around anymore 03:57
I'll see you ~9hr from now 03:58
good night and thanks for chasing this down :)
svnbot6 r14685 | Coleoid++ | Fixes to navigate around _svn dirs when svn is configured to use those instead of the usual .svn folders. 03:59
Coleoid Just enough help--thanks! 04:00
04:01 weinig is now known as weinig|zZz 04:04 mbradley is now known as mbradley|phone
svnbot6 r14686 | Coleoid++ | Slight delay as I found an editor that would not mangle the Unicode in AUTHORS 04:15
04:22 mbradley|phone is now known as mbradley 04:30 dduncan joined 04:34 fayland__ joined, fayland__ is now known as fayland
GeJ win 2 04:39
05:01 folli joined 05:16 SubStack joined 05:47 BooK joined 05:48 scook0 joined, ivas joined 06:00 BooK_ joined 06:10 mbradley is now known as mbradley|zzz 06:13 DHGE joined 06:15 BooK joined 06:17 johnSt joined
johnSt audreyt: please do not be rude to chromatic -- why won't you work on what he suggests you should be working on? let's cooperate to make perl 6 the greatest language! 06:19
06:21 zgh joined 06:38 DHGE joined
johnSt can anyone explain why the concerns of chromatic (perl.com editor) have not been addressed? he is perfectly positioned to survey the perl landscape, yet it seems like no one is following his command. we harm only ourselves by ignoring the natural leaders in our community imho. 06:48
06:49 scook0 joined
zptao natural leaders <--- lol 06:57
06:58 johnSt left
wolverian heh. 07:02
dduncan yeah really 07:09
07:35 bsb left 07:44 iblechbot joined 07:45 chromatic joined
chromatic I'm not trying to be a natural leader or anythiing. 07:45
I have some experience, yeah.
I'm just some guy though, and I can be a real jerk sometimes.
If I'm wrong, that's fine. I'm wrong sometimes. 07:46
If I'm not helpful, then somebody with some gusto just grab me in private and set me straight. Please. 07:47
But yeah, I get frustrated sometimes and I really hate when there's something wrong and a lot of people think it and none of them want to say it. That makes me want to say it.
Again, though, I'm wrong sometimes.
I'm a jerk sometimes. Sometimes I can even tell before I say something.
If I'm right, it's not because I edit Perl.com or write books sometimes or have written some software that some people use. 07:48
I'd like to think that anyone could say something that's true and have people consider it seriously. 07:49
Likewise, if I'm wrong, I hope that people are willing to call me on it.
I've sent fglock and audreyt an email outlining what I think needs to be in place before I can seriously look at making a PIR emitter for MiniPerl6. 07:53
dduncan personally, I don't know what the fuss is about ... all I saw you say is you were disappointed in the progress of Perl 6 and/or how easy it is to use the implementations
nothing to worry about I think
that is, I don't think there are any problems
chromatic I saw the backlog where someone said something like "He's the editor of Perl.com, so you guys should take him more seriously!"
dduncan I saw that too, and don't know what he's referring to 07:54
I looked further backlog, and on several web sites,
I saw no disrespect between you and audrey
chromatic Thanks for the perspective.
dduncan and fyi, I know that other people I encounter are also somewhat pessimistic about the progress of Perl 6, though I am personally optimistic about it 07:55
I think its moving along well, even if lots of others don't see that
but I also haven't really tried to use it yet ... 07:56
or I have, and decided it will probably be ready later
and am continuing in Perl 5 for the moment
now that the Metamodel has just been attached to the Haskell, or is in progress, thats a big step forward
chromatic Yeah, Larry pointed that out to me too. 07:57
dduncan and was one of the main things I was waiting for
fyi, I'm trying to implement an integrated RDBMS in Perl 6, and the Perl 6 introspection abilities were a prerequisite
chromatic I wanted to use Test::Class in Perl 6, but OO support was a prerequisite. 07:58
dduncan at the moment, I'm making a Perl 5 prototype first, then doing it in Perl 6 after
we've had fake OO of sorts for awhile, but the new metamodel should make it more real
chromatic FWIW, having a decent OO language for TGE would help Parrot and its hosted languages move much faster too. 07:59
dduncan now, afaik, we're still waiting for GHC 6.6 to be ported to some platforms, then 6.4 cruft can be removed from Pugs
I decided awhile back that, while Parrot is to be the long term best platform for Perl 6, I would be using the Perl 5 backend in production first 08:00
chromatic There's where we start to disagree.
08:01 buubot joined
dduncan well, if Parrot + Perl 6 work in production in the next few months, let me know 08:01
but I guess TGE is part of making that so
chromatic That's honestly my big block right now, technically. 08:02
As well as that of Perl 6 on Parrot.
dduncan fyi, my thought of Perl 5 working first was based on the idea of Perl 5 being stable while Parrot is pre-alpha ... while I'm not too close to the Parrot side of things, I was under the impression that other Perl 6 backends were moving along a lot faster than it was 08:03
chromatic As VMs, yes, that's likely true (unless you count YARV).
dduncan I saw the Parrot people having a more difficult task, hence it taking longer
chromatic As in supporting the features that Perl 6 needs? I don't have that impression. 08:04
I get this sense of a looming Turing Tar-Pit for some of them, especially Perl 5. 08:05
dduncan fyi, another part of the backlog shows Audrey is disappointed in having missed you in this channel, like she wants to get you involved some more ... but I guess your statement of prerequisites is your lead-in to that 08:06
chromatic Yeah. I would like to make the best possible use of my time. 08:07
dduncan now, her last post was that she'll be back online in about 9 hours ... that was made about 4 hours ago 08:08
what TZ are you in?
chromatic -0700 I believe
dduncan and I'm -0800 ... you probably won't be up then
chromatic Not if I can help it! 08:09
nopaste? 08:10
dduncan fyi, if it explains anything, most of the short term value I want to get out of Perl 6 is the syntax, the fact everything is an object, stronger types, sets, etc ... essentially, a cleaner Perl 5 with less verbose syntax for common operations ... a lot of the big ticket features I still have to learn to find a situation where I would use them 08:12
oh yes, and the built-in powerful grammar
chromatic Once you go multi, everything else seems so unflavored. 08:13
pasteling "chromatic" at 63.105.17.30 pasted "Make and Test Times on Fresh Checkout (#14685)" (34 lines, 1.5K) at sial.org/pbot/21006
dduncan there it is
but normally its instantaneous
chromatic The build or the test? 08:16
dduncan I was referring to your nopaste question 08:17
a submitted nopaste tends to appear instantaneously in the channel
chromatic Oh. No, I was trying to get a bot to tell me the URL.
dduncan I just bookmarked the nopaste in my browser long ago
?nopaste 08:18
lambdabot paste.lisp.org/new/haskell
dduncan I thought, ?eval had the questionmark in front, so maybe that too
chromatic Good thinking.
dduncan but that url doesn't work 08:19
I bookmarked sial.org/pbot/perl6
lambdabot Title: sial.org Pastebot - pasteling
chromatic I just tickled pasteling until it gave up user info. 08:20
08:20 fayland__ joined
dduncan now, since Perl 5 is more stable than Perl 6, I have been experimenting in making the RDBMS there first ... while Perl 5 is more verbose, it is still good for demonstating what semantics I actually want, so it is easier to get appropriate help in making that work in Perl 6 08:21
my code in progress is at darrenduncan.net/QDRDBMS/ 08:22
lambdabot Title: Index of /QDRDBMS
Cugel Just a quick Perl5 question for my personal education. I'm using " use lib "."; " to force Perl to look in my current dir for my .pm file (all other methods seem to fail for some reason). Is this a hack or is it Good Coding?
dduncan you shouldn't need to
chromatic Depends on why your .pm file is in your current directory.
dduncan Perl 5 always looks in the current directory, AFAIK 08:23
chromatic Not with taint mode, I believe.
dduncan but keep in mind that, if you are invoking your script from a soft reference, the CWD may not be where your script is
as is often the case with web apps
Cugel Yeah for some reason even updating @INC doesn't seem to work. 08:24
dduncan for security and other reasons, I always locate my web code in some other dir, and just put a softlink to that in the httpdocs
Cugel I see.
dduncan but that also means the script needs to have an explicit use lib in it so Perl finds the other parts
Cugel I'm finally getting the hang of variable scoping and using strict and warnings etc. I'm still the BASIC guy I was 20 years ago apparently. 08:25
08:25 SCalimlim joined
dduncan so try adding a "use lib '/path/to/dir/holding/your/modles';" to your main script and see if that helps 08:25
my paths are usually absolute in that situation
Cugel good idea.
dduncan I habitually put this at the top of all my Perl 5 code: use 5.008001; use utf8; use strict; use warnings FATAL => 'all'; 08:26
making warnings FATAL is particularly useful for web scripts 08:27
Cugel So each warning is not just a warning but a real error then?
dduncan since normal warnings won't turn into exceptions that I can catch and print to the browser for debugging
a normal warning isn't an exception; I make them into such 08:28
Cugel I'm copying that, looks good.
08:28 elmex joined
dduncan my main script is a big eval block which contains everything else ... that eval block would only catch an exception if something in my code dies 08:28
jrockway why the use utf8? do you use a lot of unicode literals in your scripts?
dduncan that is, the main script contains almost nothing but the eval block ... all my real code is in modules 08:29
I use utf8 by default, so just in case I have non-ascii chars, they are unicode and not some other encoding
I make my whole pipeline unicode
Cugel What does the use 5.008001 do really? 08:30
dduncan it specifies minimum perl version to run under
you can set it different if you want, but the above is the first useable 5.8, and I don't want older than 5.8
Cugel I'm not using OOP yet, so my minimum version might as well be Perl 5.0 :-) but still.
dduncan if you have specific needs, you can set it differently
you could leave the line out entirely, as well 08:31
jrockway use utf8 means "my source file is encoded in utf8"
dduncan but fyi, the utf8 pragma only works with 5.8
chromatic You need at least 5.6 if you want the warnings pragma too.
dduncan as well as some other things I want to use
jrockway if you want data read from files to have the utf8 flag on, you need to use other means
dduncan yes, the pragma just says the Perl code is utf8, not file contents etc
but 5.8 also has PerlIO and stuff, so its easy to tell it what encoding to treat files as 08:32
as well as standard out and stuff
Cugel Ah I didn't know 'use warnings' was relatively new.
jrockway yeah, i don't see any reason to support perls that are 5+ years old
if your OS doesn't ship with perl 5.8.x, it's time to get a real OS 08:33
dduncan or install it yourself
jrockway yeah; wanted to avoid the "perl is too hard to compile" discussion :)
dduncan I run Mac OS X, and 10.3+ include 5.8
but I still compile it myself 08:34
jrockway IIRC apple picks weird options from time to time
dduncan how so?
I know their bundled perl is multithreaded, while the default Perl config is not 08:35
jrockway i think in the 10.3 days they didn't include ithreads, the included 5.0005 threads, or something
dduncan I don't know
I've never used any multithread features
jrockway after 10.4 the included perl was fine
08:35 gene9 joined
dduncan I'm on 10.4.8 now 08:35
jrockway hmm, my last memory of OS X was 10.4.5, which hosed all my production servers 08:36
i don't remember why, though...
apple's take on apache was pretty weird 08:37
ld wanted to build all libraries as "universal binaries", but apache didn't appreciate that
but apple changed the source of ld to to ignore -arch flags, or something
so i had to build my own universal apache
that was a nightmare
pasteling "duncand" at 24.69.53.198 pasted "example (perl 5) web script wrapper" (26 lines, 724B) at sial.org/pbot/21008 08:38
dduncan Cugel, that may help you 08:39
its a generified version of my "main program" for a web app
jrockway at that point, why not use CGI::Application or Catalyst? 08:40
dduncan it just invokes the real main, and traps and reports any problem with either compiling or executing such that the real main doesn't handle itself
jrockway the catalyst error screen is prettier ;)
Cugel Looks good -- very educational too.
dduncan at the moment, I just wanted something with as few dependencies as possible ... once I actually have experience using those frameworks and trust them, I'll switch over, though 08:41
jrockway btw, what happens if $err contains HTML?
i don't see the content-type set to text anywhere
Cugel I love the 1; at the end -- is something like that still in Perl 6? 08:42
jrockway there's a good reason for that
dduncan an omission ... but it matters little since that trap should never show up but when I'm actually editing the site and made a typo 08:43
jrockway you could do "return _init" or something, and if _init fails the module won't load
dduncan actually, the trap I showed you is mainly to catch COMPILE errors of the program
jrockway ah, i see
if haven't done plain (perl) CGI for a while
dduncan the 1; at the end isn't needed for Perl 6
it isn't always needed in Perl 5 either, but it doesn't hurt to have it, to be safe 08:44
fyi, I consider CGI cludgey, but only used it for a quick solution ... my program has its own "IOUtils" module which completely wraps it anyway, so the program doesn't suffer its API 08:45
fyi, part of the reason for not having used Catalyst et al, besides inexperience, is because the script in question started out as a PHP one that I inherited, then translated 08:46
the $job said first to just get it working like the old one did sortof, then we can later do a better rewrite 08:47
jrockway yeah, understandable 08:48
i am in the reverse situation now -- moving old perl to "new" PHP
Now With Less Maintainability (tm)
don't get me started on PHP, though
dduncan I agree ... PHP is teh suck
jrockway it really impresses me how bad of a job the PHP devs did
dduncan that said, I'd sooner use it than some MS tool
jrockway they just don't understand programming, but they somehow managed to write their own programming language 08:49
i'm not sure what to think about that
i also love their helpful default-bison parser... "parse error: syntax error: expecting T_FOOBAR but got T_STRING"
that's helpful.
but yeah, not going to go off on a PHP rant ;) 08:50
must... resist...
pasteling "folli" at 200.161.234.112 pasted "teste" (1 line, 5B) at sial.org/pbot/21010 08:57
dduncan jrockway, I should also point out, in my code example, that $err is unlikely to ever contain HTML ... usually it is either an error generated by Perl (terminal output intended), or it is a SQL statement
so the HTML mode will word wrap so it displays on the screen at once 08:58
jrockway for production it's a good idea to entity-encode everything your print to the browser, just to be safe
unless it's literal HTML, of course
then you just have to be careful
dduncan I do entity-encode all normal output, as well as any errors reported by the main program itself, which has another eval block over everything ... the one I showed you is mainly just for when that main doesn't compile 08:59
fyi, when I translated from PHP, I replaced all ?>...<?php with print q{...}; 09:00
which is what that literally means in PHP
my text editor supports regexps, so it was fairly painless 09:02
eg, a lot of the conversion was search'n'replace
09:16 drrho joined 09:18 buu joined 09:23 buubot joined 09:27 buubot joined 09:40 andara joined 09:45 buubot joined 09:52 kensanata joined, buetow joined, nwc10_ left 10:02 buubot joined 10:11 spo0nman joined 10:14 fglock joined 10:17 eggzeck joined 10:22 ruz joined 11:05 ruoso joined 11:09 dduncan left 11:14 xinming joined 11:18 Odin-LAP joined 11:54 pancake joined, pancake left 11:55 ofer1 joined
dmq TimToady are you about? 12:06
Steve_p dmq, it'd be pretty early for California time 12:07
kane-xs_ 4 am ;) 12:08
dmq aha 12:09
silly me. :-)
these crazy californians sleeping in so late. geeze. :-) 12:10
12:11 chris2 joined 12:12 Psyche^ joined 12:16 BooK_ joined 12:28 Psyche^ is now known as Patterner 12:30 electrogeek_off is now known as electrogeek 12:38 elmex joined
audreyt @tell johnSt I try, sometimes very hard, to not be rude to anyone, chromatic included. You may take comfort in knowing that the concerns he previous raised, namely "make test" taking 5hr on his machine, was the main reason of the refos on performance in 6.2.x... 12:41
lambdabot Consider it noted.
svnbot6 r14687 | fglock++ | MP6 - fixed 'method', add tests
fglock audreyt: ping 12:42
lambdabot fglock: You have 1 new message. '/msg lambdabot @messages' to read it.
audreyt fglock: pong 12:45
fglock audreyt: I'll start working on the operations parser 12:47
after that, it should be able to compile Token->P5
which means MP6->P5 :) 12:48
audreyt ok. sorry I was disabled thoroughly yesterday :)
fglock are you enjoying Rio?
audreyt yes 12:49
sugarloaf and the copacablanca beach is nice indeed
did quite some design work on paper on MO integration during Rio.pm meeting 12:50
is your new laptop relatively stable now after reinstall? 12:52
fglock yes, it is fine now - which font do you use for terminal?
dmq wingdings 12:53
fglock :)
dmq keeps the management on their toes
audreyt fglock: Consolas 12:55
dmq: but it's not monospace...
fglock audreyt: try out t/04-regex.t - although it still breaks in the end of compilation
dmq sorry i meant wingding-mono 12:56
12:56 Limbic_Region joined
dmq of course wingding-mono has the disadvantage of only having one char in it. 12:57
:-)
audreyt rofl :) 12:58
and it's a space
hence the name "monospace"!
it's a very bleachy clean font, it appears
fglock: k, trying
dmq Indeed 12:59
audreyt fglock: the most complex part of emitter is probably the Bind node
fglock: since it needs to handle array and object unpacking
but even that is ~20 lines I think
just need a different emit metod for the relevant notes in lvalue position 13:00
fglock: Undefined subroutine &Pugs::Runtime::Perl6::Scalar::join called at lib/MiniPerl6/Emitter.pm line 379.
normal?
fglock audreyt: how about emit Bind->MP6
audreyt: may need to reinstall v6 13:01
audreyt you mean simplifier?
i.e. desigar
desugar
fglock audreyt: yes
audreyt [$x,$y] := ...
into...
$tmp := ...
$x := $tmp[0]
$y := $tmp[1]
?
sure, that is good thinking 13:02
fglock yes - this would simplify backend implementation
audreyt otoh, some backend may gen more efficient code having acces to full bind node
so how about providing a .simplify method
to Bind, and backends can choose to run it or not?
fglock I'm thinking of a generalization 13:05
we have a few DSLs - regex, string interpolation
these could be "optional" too 13:06
13:06 DgBR joined
fglock so rich backends can be more efficient 13:06
audreyt you mean regex becomes simplified similarily 13:12
so you call .simplify when you want a recompiled inline
and handle it yourself if you want
that sounds like sane
fglock yes - but we need to find out how to implement this without inheritance 13:13
also, how to handle multiple backends
audreyt each backend is one class of Emitter
you don't need inheritance; when you're being blessed there's an attribute of the generalized emitter-helper object 13:14
i.e. use hasa to emulate isa
unfortunately my blood sugar level is very low at this point
let me hunt for some glucose and then come back 13:15
in say 1hr?
fglock np - better finish the base grammar first, also
ok - I'll have lunch in 30min, so after lunch
audreyt ok
fglock did you get t/04-regex working? 13:16
audreyt }, 'Var' ) 13:17
]
*** error in Block
is the output. it parses "my" as function call
not as decl
but otherwise seems to be fine
fglock yes, but it goes very far already :)
audreyt it does! esp. considering the timeframe :) 13:18
fglock unfortunately we didn't finish in 1 week - let's try in 10 days 13:19
13:20 iblechbot joined
fglock bbiab 13:20
13:27 bpphillips joined 13:29 eggzeck joined 13:30 martynoff joined
smash_ good afternoon 13:32
13:37 boo_ joined 13:43 GnuVince joined 13:57 bonesss joined 14:06 BooK joined
audreyt fglock: sorry - the restaurant wasn't open until now 14:10
fglock: so I'll be back later than that -- also my brain chemistry is not in a properly hacking confirmation at this moment -- I'll try to do something to make myself feeling better about working on Pugs
s/confirmation/configuration/ 14:11
so, bbiab
Gothmog_ audreyt: On pugscode.org the link to the svn repository is still the old one... 14:17
fglock audreyt: food++ 14:20
stevan fglock: is there any docs about what MP6 looks like? 14:31
and would it make sense to try and write the metamodel in it (similar to what I did in PIL a while ago)
fglock stevan: that would be great 14:34
stevan: we probably still need native metamodels
but having an easily deployed one is nice
stevan fglock: by native metamodels, you mean like MOH and Moose?
native on the backend 14:35
fglock see /docs/notes/miniperl6-bootstrap.pod
yes, native in the backend - because it can be faster
stevan ah 14:36
ok
fglock the plan for MP6 version 2 is to include a metamodel in the base language
stevan what does v1 do? 14:37
fglock if we have a metamodel written in MP6, we can use MP6 to write MP6-2
stevan circularity++ :) 14:38
fglock v1 does closure, tree transformations
it also has a parser 14:39
14:39 beu is now known as beuster
fglock miniperl6-bootstrap.pod is an overview; pugs.blogs.com/photos/visiolization...strap.html is about the inner workings 14:41
lambdabot Title: Visiolization: Mp6bootstrap, tinyurl.com/yxb6ux
fglock v6/v6-MiniPerl6/AST.pod and BOOTSTRAPPING are more detailed internals
stevan nice 14:42
fglock It already works; I'm currently adding additional nodes in order to complete bootstrapping
kolibrie stevan: did you notice there might be an early May hackathon in Boston? (discussed yesterday) 14:43
stevan kolibrie: no, I didnt
fglock: where is v6/ 14:44
kolibrie clkao will be in the area, so audreyt thought of coming too, and cj was sort of arranging things
stevan oh I see it
kolibrie: cool, I will swing on up then 14:45
fglock it's in pugs/v6/v6-MiniPerl6 - you need a fresh install ov perl5/Pugs-Compiler-Rule and -Perl6
kolibrie I talked with my wife about it, so I may be able to come
fglock of
stevan kolibrie: cool 14:46
fglock it needs Pugs::Compiler::Perl6 because the bootstrap is not finished
it may compile with pugs too - I haven't tried 14:47
14:49 vel joined
fglock stevan: there is a lot of sample code already written - MP6 is implemented in MP6 14:52
stevan fglock: it looks like you already have objects 14:58
14:58 rlb3 joined
fglock stevan: yes, but it's a very simple system - it does not support inheritance 15:00
it's mostly just sugar 15:01
stevan fglock: structs basically?
fglock it's blessed structs 15:02
stevan thats perfect,.. that was all PIL has really
fglock some evil soul may evan want to emit it to C...
svnbot6 r14688 | audreyt++ | * Add pugscode.org into version control.
fglock even 15:03
MiniPerl6 is PIL-3
stevan fglock: cool :)
svnbot6 r14689 | fglock++ | MP6 - added infix expressions 15:05
15:08 weinig|zZz is now known as weinig|bbl, cjeris joined
weinig|bbl 15:12
sorry
audreyt Gothmog_: thanks, fixed; please feel free to update the pugscode.org homepage as you see fit 15:16
by simply committing to docs/feather/pugscode.org
bbiab again... 15:17
svnbot6 r14690 | fglock++ | MP6 - implemented ternary; substr() works too 15:21
15:23 fglock joined
fglock the main language impl is done - now bugfixing... 15:24
if I copy the .svk/ and pugs/ directories from another computer, would it work? 15:27
[particle] .svn dirs? 15:28
15:28 hexmode joined
[particle] i have moved an svn working copy between dirs on the same machine and it works. dunno about svk, and haven't moved between machines. 15:29
PerlJam fglock: probably. (I've never done it either, but my gut says it should work) 15:30
[particle] fglock: can you write up detailed steps to get MP6 running? stick them in a README or something? i'd really like to play, but i can't get it to work (same problem as chromatic, in fact)
fglock [particle]: it will be easier after bootstrap 15:31
[particle] eta?
fglock right now, you need an up-to date v6.pm
2 hours :)
[particle] -I path/to/v6.pm doesn't seem to work 15:32
ok, i guess i can wait
fglock just kidding - but I hope to finish today, with audreyt++ help
[particle] ok
i'll be on a plane in 4 hours
er, 5
fglock can you get v6.pm to run?
[particle] how? 15:33
fglock cpan v6 - then update from perl5/Pugs-Compiler-Rule and Pugs-Compiler-Perl6 - this would give you an up-to-date installation
15:33 justatheory joined
fglock which can run MiniPerl6 15:34
[particle] ok, v6 is fine from cpan
should i expect failing tests?
fglock yes 15:35
many, indeed, as we've been changing the implementation in order to just get MP6 working
[particle] ok, make seems to be taking a while... done. 15:36
fglock just perl Makefile.PL, make, make install
[particle] PCR done 15:37
PCP6 done
fglock then go to v6/v6-MiniPerl6 and try perl -Ilib t/02-exp.t - this should show you an AST and some emitted perl5 snippets - that's the compiled code
no TAP tests yet 15:38
[particle] 02-emit.t
D:\usr\local\pugs\v6\v6-MiniPerl6>perl -Ilib t\02-emit.t 15:39
D:\usr\local\pugs\v6\v6-MiniPerl6>
fglock one sec, adding t/02-exp.t
[particle] at least now i'm not getting errors
oh, i may not be at HEAD
upping
15:41 pmichaud_ is now known as pmichaud
svnbot6 r14691 | fglock++ | MP6 - added t/02-exp.t 15:41
fglock t/02-emit.t works here 15:42
[particle] no output
fglock try: perl -Ilib t/02-emit.tc
[particle] D:\usr\local\pugs\v6\v6-MiniPerl6>perl -Ilib t\02-emit.tc 15:43
syntax error at lib/MiniPerl6/Grammar/Regex.pm line 4249, near "};"
syntax error at lib/MiniPerl6/Grammar/Regex.pm line 4251, near ","
Compilation failed in require at lib/MiniPerl6/Grammar.pm line 20.
BEGIN failed--compilation aborted at lib/MiniPerl6/Grammar.pm line 20.
Compilation failed in require at t\02-emit.tc line 19.
BEGIN failed--compilation aborted at t\02-emit.tc line 19.
fglock looks like a PCP6 error
if you upgrade PCP6, you then need to remove the bad .pmc files in lib/ before trying again 15:46
[particle] ok, which lib? 15:47
mp6?
fglock yes
sorry about the teething problems 15:48
[particle] i thought that happened automatically (recreating .pmcs) 15:49
fglock it only recreates when you touch the source file; but in this case it's a compiler error, which it can't detect
[particle] ok, same error, different line
syntax error at lib/MiniPerl6/Grammar/Regex.pm line 4413, near ";" 15:50
etc
fglock one sec - let me clean up and try
15:53 Limbic_Region joined
fglock it still works for me - though now I get an error in "Class/MOP/Instance.pm during global destruction" 15:54
better wait for MP6 bootstrap :(
[particle] will do, thanks for holding my hand 15:55
fglock np - PCP6 has some pretty complex module interactions stuff 15:56
16:12 nothingmuch joined
nothingmuch sup homies 16:18
[particle] moose, yo 16:20
fglock nothingmuch: hi 16:22
svnbot6 r14692 | fglock++ | MP6 - parenthesis, &&, || 16:23
16:24 DHGE joined
fglock meeting & 16:30
16:36 nothingmuch_ joined 16:39 weinig|bbl is now known as weinig 16:41 bpphillips joined 16:49 buetow joined 16:56 araujo joined
GabrielVieira fglock hey 17:09
17:16 weinig is now known as weinig|bbl 17:47 kanru joined
fglock GabrielVieira: pong 17:50
GabrielVieira :) 17:56
fglock ola
17:56 andara left 17:57 thepler joined
svnbot6 r14693 | fglock++ | MP6 - fixed $obj.meth 18:05
18:07 avar joined 18:12 zgh joined 18:20 CUTandRUN joined 18:39 Grrrr joined 18:50 bernhard joined 18:54 justatheory joined 18:59 nothingmuch joined 19:06 thepler joined, kanru joined, boo_ joined, buubot joined, badalex joined, lisppaste3 joined, autark_ joined, scw joined, baest joined, Eidolos joined, bcorn_ joined, jrockway joined, silug joined, pasteling joined, Gothmog_ joined, smash_ joined 19:09 buubot joined 19:12 justatheory joined 19:13 ludan joined 19:38 weinig|bbl is now known as weinig 19:52 larsen_ joined 20:07 vel joined 20:09 vel joined 20:23 justatheory_ joined
svnbot6 r14694 | fglock++ | MP6 - a bit improved expression parser 20:30
20:30 fglock left 21:03 cmarcelo joined 21:15 cmarcelo_ joined, cmarcelo joined
cmarcelo is here a good place to report typos in Syn? (or learn that they aren't typos after all) or p6-lang is the proper place? 21:29
21:29 weinig is now known as weinig|bbl 21:38 iblechbot joined
cmarcelo i'll put them here in case someone wakes up: S12.pod line 608 missing a "," in arguments for constructor? S12 line 470 begins "C<has $.foo>", shouldn't it be "C<has &.foo>"? 21:42
21:48 justatheory joined 21:52 justatheory joined 21:59 DHGE joined 22:00 bpphillips left 22:30 Limbic_Region joined 22:38 cmeyer joined 22:51 dduncan joined 22:59 weinig|bbl is now known as weinig 23:09 justatheory_ joined
nothingmuch woot bork 23:11
jrockway anyone in there that's going to be at the hackathon tomorrow? 23:17
s/there/here/
:)
thepler yes 23:20
23:20 cmarcelo joined
thepler looking forward to it 23:21
TreyHarris is something wrong with the password-changing server? i just got back from my trip, saw the notification from audrey, and went to change my password with the URL i got in her mail, but can't connect 23:24
svn.pugscode.org doesn't seem to be responding on port 9999 at all
dduncan I could get in to the new server yesterday
looking ... 23:25
I can still connect to 9999 now
TreyHarris how odd
it's not a browser issue, neither safari nor firefox nor opera work 23:26
dduncan so are you saying svn.pugscode.org:9999/ doesn't show anything for you?
lambdabot Title: Welcome to CommitBit for PugsCode
dduncan the bot can get in
Trey, perhaps you are behind a firewall blocking access to external ports outside a preapproved list? 23:27
TreyHarris dduncan: ahhh. yes, that's exactly right. thanks.
guess i'll wait till i can connect from somewhere else
dduncan unless you can contact whom admins the firewall to give you an exception
er, give that server an exception 23:28
TreyHarris dduncan: getting a port access request approved here is usually a multi-week deal :)
might as well just wait till i get home
23:28 cjeris left
TreyHarris ahh, works fine on links, so i just ssh'd elsewhere :) 23:30
GeJ win 2 23:33
damn thos morning fingers
morning all
Patterner Moin moin 23:34
GeJ Servus Servus 23:36
23:39 mako132 joined
ingy hola 23:42
jrockway hugs Alt-2 :)
nothingmuch hola ingy-san 23:46
ingy nothingmuch! 23:49
where are you?
23:51 eggzeck joined 23:55 anatolyv joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!