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 | dev.pugscode.org/
Set by putter on 11 February 2007.
putter pugs, which has the only full one (kp6 being a runner up), wont share all its info (particularly class decls). 00:00
audreyt putter: perl6::Bar is fine.
lambdabot audreyt: You have 4 new messages. '/msg lambdabot @messages' to read them.
putter which has been a dev bottleneck for... just about two years now.
re fine, :)
bottleneck might be gone by itself real soon now.
audreyt re class decls, that's because not pugs itself know about it
(no introspection. aka, no MO, hence no MO dump) 00:01
putter but doing an opp and grammar based might be good practice, drive the spec, and make 6.err, the one after next, easier to do.
audreyt re 6.283, yes. critical, even.
putter re itself, indeed :)
audreyt re MO hacking, will do that as long as disk0s2: 0xe0030005 (UNDEFINED). goes away 00:02
need to spend ~2hr running ddrescue
since it's creeped into ~/.svk/ area and hence interfere with hacking
*as soon as
putter so, given the perhaps questionable insanity of "let's write a new p6 parser", a next question is
re disk, eep 00:03
lol
audreyt mm "questional insanity" :) anyway, I'll resurface after ddrescue is done
meanwhile enjoy :) &
putter (image of audreyt watching disk progressively die, but thinking "no, it hasn't hit /svk yet, so it's not important", and goes back to hacking...;) 00:04
thus the lol
good luck
so given an attempt to write a full p6 parser, a next question is how to test. to judge progress, validated design decisions, determine success, etc. 00:05
While you need to track sub arity to do a basic parse (basic as in no BEGIN, etc), I think that's about it. 00:06
So rather than trying to create a runtime to go with the parser, one could output PIL instead. 00:07
Could check it against pugs -CPIL output, but that kind of game tends to get messy, as non-essential details complicate comparison.
But "PIL2JS passes test using new parser's PIL instead of pugs's parser's PIL" is a good proxy for "your new parser is working". 00:08
a new parser could certainly be written in haskell. should eventually (or perhaps just p6). 00:17
isnt p6 now because I don't trust pugs (without audreyt at hand) to not have gotcha path blocking bugs.
but hs vs p5, the regex engine I have the greatest short-term hope for is p5. hmm, but with p5 now linked into pugs, that's mostly a non-argument. 00:19
part of the p5 vs hs thought is that the p5-based components of pugs seem to have drawn more developers and developed comunities more than the hs component has. 00:20
eg Moose
end of first level response. ;) 00:21
putter s/-CPIL/-CPIL1/ -CPIL1-PERL5 being used by PIL2JS (I think?). -CPIL1-JSON or -CPIL1-YAML perhaps your best bet ‎if you want to try your hand at writing a new backend. 00:26
That means *you*. :)
(for every/and(@you)) 00:27
s/and/any
It's not an enormous task. PIL2JS and PIL-Run were sortof working in just a day or three. With PIL2JS to serve as a prototype, if you keep the same design (a good idea), its basically a transliteration exercise. 00:29
end of advertisement
cd perl5/PIL2JS/; ./runjs.pl -e 'say "Hello, World!"' ... doesn't say anything. :( 00:33
@seen iblech
lambdabot I haven't seen iblech.
putter hmm... very old spidermonkey 00:42
Gothmog_ 0 00:48
Whoops.
gaal putter: you moosed? 00:57
I'm back from a few drinks w/ friends from around and abroad so I'm thinking as straight as my antlers but... elk? 00:58
putter lol 01:00
laugh, groan, etc
gaal was funny, but we were just looking for a place and accidentally landed on a pick-up bar so was v loud and smokey etc. 01:01
I'm about to hit the shower and sleep
xerox Hit the shower, hit the sack. Two frags in a row. 01:02
putter I was going to ask about the new tree, and where to find it. but I decided the old tree in the back yard would make a better subject, since I could more easily compare it to the new parser's tree sketches.
gaal Pugs.Val
and Pugs.Val.*
putter danke 01:03
gaal some of it is active already
also some possibly outdated "rest of it, not yet transcrivbed" is in CaptInternals.hs
s/t// maybe
however, be advised audreyt has some sekritt plan to simplify the basic scheme 01:04
putter ok. PIL2JS isnt working for me, so the old tree may not... fly doesnt seem the right word but...
gaal ie: newVal currently is based on a variant moose of IIRC 5 subtypes
native, uh something else immutable I guess, mutable (==STM), external (==IO) 01:05
putter oooo, secret simplicity plans. complexity seems to be the thing which always stops "try to take over the world" projects.
gaal the first to are represnted by rank-1 simple |variants like a sane haskell programmer would do
however as everybody knows, there's not really such a thing (!ExSx) as a sane Haskell programmer, 01:06
putter ("i didn't say it") 01:06
gaal the third three variants are existentially quantified
putter k 01:07
gaal "to be or not to be? It depends, what monad are you in, for starters?"
putter hmm.... stm, the death of alternate backends?
gaal hmm i dunno
putter lol
gaal it was a reaosnable transcription of the prexsiting, at the time, L<S06/"Mutable Types"> 01:08
putter ooo, oo, maybe if I write an arrow library for p6, noone will ever use monads in 6! oops, evil plan revield.
gaal so ask TimToady -- I don't think he was thinking specifically STM when he enumerated these types but maybe he was?
putter k 01:09
lots of delightful reading pointers. thanks! :)
gaal or oh, maybe no, regualr mutable <=> IO types is an audreyt distinction
not sure if it's even reflected in the S06 enumerations
sorry too drink to look up now 01:10
*u
putter cheers, tnx
gaal yeah that shower.... mm&
oh also another pointer... an advantage of using existentials as they are layed out in this newVal thing is "open types" 01:11
@google rafl open types
lambdabot bugs.xmms2.xmms.se/view.php?id=1031
Title: 0001031: No mime type reported for files in the medialib. - Mantis 01:12
gaal @google ralf open types
lambdabot portal.acm.org/citation.cfm?id=1140...EN=6184618
Title: Open data types and open functions
gaal (oops)
yeah, that one
the idea is code organization, you can define a type in some Hs compilation unit other than where the base Val type was mentioend
and augment it 01:13
ie., this is (also) a Hs implementation/engineering decision
ok where's that shower zz&
avar gaal: I was looking at your smoke server slides and was wondering about ripping the smoke serv client script out as Module::Install::Smoke 01:16
smokeserver '..../serv'; # In Makefile.PL 01:17
I wonder what cpan testers use, you could optionally submit stuff to them 01:18
"Do you want to upload smokes along with make test [y/n]"
avar sleep & 01:24
gaal avar: look at Chimps on CPAN and talk w/ obra -- they use this in production AFAIK 01:25
gaal putter: one last thing -- re: portability and STM 01:25
you can think at the very least of the VMut / VExt distinction as a helpful boon for non-Hs backends, like so: 01:26
condent { say "oops, I shouldn't run" }
svnbot6 r15336 | putter++ | READTOO - added new (duplicate;) entry for "Open data types and open functions". Can be deduped later. For now, keeps file active and builds interest.
gaal -- this is a runtime exception in haskell backend for free
putter listens attentively.... 01:27
gaal how do backends w/o natural STM know that say is unsafe fro STM?
TimToady *contend
gaal well, they don't, except hey, helpfully perhaps non-STM thingies in the AST are marked as VExt :)
so at the very least this is useful documentation 01:28
avar gaal: ohh, awesome 01:28
putter ah
TimToady and the thread continuation monad may provide similar for thread vs event issues
assuming we take the approach of www.seas.upenn.edu/~lipeng/homepage/unify.html 01:29
lambdabot Title: Unifying events and threads
gaal in short (you folks)++ and I shall now take my leave etc.
& fer real
putter back. thanks gaal! 01:30
TimToady someone else mentioned it first a while back here... 01:35
svnbot6 r15337 | putter++ | READTOO - added "A language-based approach to unifying events and threads".
putter ?
TimToady hmm, must have been somewhere else I saw it first...colabti only sees your new entry. 01:39
maybe it was in the chatterbox 01:40
allbery_b it was mentioned here, IIRC
TimToady yeah, audreyt, colabti.de/irclogger/irclogger_log/...9,Fri#l363 01:43
lambdabot Title: #perl6 2007-02-09,Fri, tinyurl.com/243zsh
putter ok. I notice in READTOO's log masak correcting "credit" and date of mention. err, it's just a slush file created because I was tired of having to grovel back through colabti. the date was just included for staleness, and the who, I don't know, randomness. just in case you wanted to discuss it. 01:47
svnbot6 r15338 | lwall++ | correctly ascribe "unify" first mention to audreyt
r15338 | lwall++ | some random cleanups in P6STD
putter bemused, putter goes back to hacking 01:48
though, perhaps the thing is to *add* mentioned by lines, each time its mentioned, and any extra links (and dates?) 01:49
@audreyt Was test-js actually working for you back around -r14356? It happily runs for me, but has 01-sanity failures. :/ 01:58
lambdabot Unknown command, try @list
putter @tell audreyt Was test-js actually working for you back around -r14356? It happily runs for me, but has 01-sanity failures. :/
lambdabot Consider it noted.
putter fixing PIL2JS is already a ticket. 02:02
probably shouldnt have spent good hacking time on it, but if it had worked, that would have been fun. 02:04
there is a call PIIL2JS from p5 with callbacks embedding existent. so could easily set PIL2JS up to accept PIL trees created by new p5 code. 02:05
could get "say 3" parsed, ast built, PIL emitted, and run on PIL2JS, in no time. ;) 02:06
good night all & 02:12
svnbot6 r15339 | lwall++ | P6STD syntax cleanups 03:03
devbot6 planet6: Audrey Tang: ddrescue'ing <pugs.blogs.com/pugs/2007/02/ddrescueing.html> 04:04
devbot6 planet6: Luke Palmer: Dwimmery sanity <use.perl.org/~luqui/journal/32489?from=rss> 04:42
svnbot6 r15340 | lwall++ | Getting P6std runnable by pugs by postprocessing. 05:21
svnbot6 r15341 | lwall++ | cheat Perl-6.0.0-STD.pm | pugs -CPIL1-YAML now succeeds 05:42
TimToady well, it succeeds in the sense that it doesn't blow up... 05:49
There are still lots of infelicites, such as the fact that only one of each multi regex shows up...
which is likely something the "cheat" script will have to deal with anyway for bootstrapping. 05:50
rindolf Hi all. 09:00
lambdabot rindolf: You have 1 new message. '/msg lambdabot @messages' to read it.
rindolf moritz: here? 09:30
moritz rindolf: yes, now ;) 09:59
rindolf moritz: I fixed the broken link. 10:01
moritz rindolf: ok ;) 10:02
rindolf moritz: thanks for reporting it.
moritz rindolf: np ;)
rindolf I'm nearing r1000 in my homesite's repository.
meppl gute nacht 13:19
svnbot6 r15342 | kudra++ | Content added 14:33
devbot6 planet6: Audrey Tang: Weekly Perl 6 mailing list summary for 29 January - 3 February, 2007 <pugs.blogs.com/pugs/2007/02/weekly_...1.html> 14:43
rindolf Hi all. 15:29
Limbic_Region salutations rindolf
rindolf Hi Limbic_Region
Limbic_Region: what's up?
Limbic_Region me unfortunately
I think I would rather be in bed sleeping 15:30
you?
rindolf Limbic_Region: well... 15:32
Limbic_Region: I added Google AdSense to my DocBook/XML-generated essays and to my presentations. 15:33
Limbic_Region I saw that in #perl
rindolf Limbic_Region: and prepared a presentation about the Joel Test for Python-IL.
Limbic_Region is pretty ignorant of adsense
well, it sounds like you are being productive - that's more than I can say :-) 15:34
rindolf Limbic_Region: more than you can say about yourself?
Limbic_Region correct
rindolf Limbic_Region: OK. 15:35
dmq juerd? 15:36
summon juerd 15:38
:-)
putter @tell audreyt Could the pugs backend be driven by externally generated yaml? Ie, do you have a feel for how fragile it is/isn't? How much implicit/idiomatic dependency is present. I was thinking of exposing loadPreludePC... 15:54
lambdabot Consider it noted.
SamB putter: what do you suppose is the best way to find out? 15:55
btw, yaml seems fiendishly complicated...
putter Ask audreyt. :)
SamB was thinking "try it"
moritz was thinking "let somebody else do the work" ;) 15:57
putter re "someone else", yes, delegate questions to someone who can answer without thinking hard, rather than spending days generating a not-as-good answer. :) 15:59
re trying it, let's see...
moritz putter: I fully agree ;)
putter Hmm, I was thinking it non-trivial to try/play, but.... if one hardcoded loadPreludePC to load a second file in addition to the prelude, if it worked at all, one could then dump something simple in the file and see if it showed up.... 16:02
(mutating or replacing the existing prelude appearing much more problematic alternatives) 16:03
SamB++
(though it's still mission creep :( :) 16:04
bbl (or tomorrow) &
rindolf Hi all. 17:39
How's audreyt ? 17:41
allbery_b recovering and back in circulation
rindolf allbery_b: nice. Good to hear.
allbery_b ...but will probably be <100% for a few months yet 17:42
rindolf What is the timezone of the nntp.perl.org messages? www.nntp.perl.org/group/perl.cpan.t...20778.html ? 17:43
lambdabot Title: PASS Test-Run-0.0104 OpenBSD.i386-openbsd 3.5 - nntp.perl.org, tinyurl.com/2sul4j
rindolf miyagawa: here? 19:41
miyagawa rindolf: pong 20:12
rindolf Hi miyagawa
miyagawa: I already asked my question in #plagger 20:13
miyagawa: but I have a different problem now.
miyagawa yes, looking at it
Juerd dmq: pong 23:14
moritz probaply ping timeout ;)
Juerd dmq: But I'll be gone real soon; my brane hurts and I'm sleepy. 23:15
afk # zz 23:20
__Ace__ AUTHORS: 23:46
" Larry Wall wrote the original version of patch." Noway? :) 23:47
jql way
or, if it's a legend, I can confirm it was legendary even 10 years ago. :)
__Ace__ yea, havent been updated in a while.. 23:48
totally bugfree then eh?
jql they only ever upgrade diff. :)
__Ace__ heh 23:49
what does that tell us?
jql the bizzare bit is by how far diff predates patch
allbery_b even remembers lwall releasing patch :) 23:51
jql I was probably in kindergarten that day 23:52

Powered by Cro and the Raku® Programming Language.

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

Thank you!