Camelia, the Perl 6 bug

IRC log for #parrot, 2008-02-25

Parrot | source cross referenced

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 svnbotl r26053 | chromatic++ | trunk:
00:00 svnbotl : [gc] Constant PMCs with custom mark() entries now always get called during the
00:00 svnbotl : mark phase of GC, even if their live flag is set.  (Nothing clears their live
00:00 svnbotl : flag during sweep, so any of their contained PObjs might not get marked during
00:00 svnbotl : the next mark phase.)
00:11 davidfetter joined #parrot
00:19 ruoso joined #parrot
00:28 kid51 Am looking for people who can give the new configuration step tests a whirl; any volunteers?
00:34 lbr joined #parrot
00:35 lbr OHAI
00:36 davidfetter HAI
00:36 davidfetter I MAED NU FRENZ
00:36 kid51 purl:  OHAI?
00:36 purl kid51: wish i knew
00:37 lbr I made a syntax error in my first lolcode program ever
00:37 kid51 purl, Ohai is a town in the Southland Region of New Zealand's South Island.
00:37 purl OK, kid51.
00:37 lbr kid51: I was amused that my tiny patch from 2004 is still causing me to get the occasional parrot mail ;)
00:37 davidfetter kid51, it's lolcode
00:38 davidfetter http://icanhascheezburger.com​/2007/09/29/i-maed-nu-frend/
00:39 kid51 lbr:  Aha!  so can you answer the question I posed?
00:39 lbr did and didn't, sorry
00:44 lbr 2004? where did that come from.  2003!
00:48 lbr kid51: just read the ticket - I'm really the wrong person to ask whether it should be closed, since the scope changed since I was least involved
00:48 lbr my opionion, 5.8.0 time, was that perls implementation is "not good enough, but maybe best we can do"
00:49 lbr (s/least/last/)
00:51 kid51 IIUC, that was Chip's last comment on the subject as well.  But he's not around to ask.
01:08 * lbr hasn't been following the soap that is parrot :)
01:22 LimbicRegion joined #parrot
01:24 Theory joined #parrot
01:45 Limbic_Region joined #parrot
01:49 kid51 Am getting 500 Server Errors at rt.perl.org.
02:10 wknight8111 joined #parrot
02:21 jdv79 soap?
02:21 purl soap is, like, Snakes on a @($*&@ Plane!~ or http://images.ucomics.com/comi​cs/wpswi/2006/wpswi060820.gif or something you don't drop in prison or http when you actually read the spec
02:49 ptc_ joined #parrot
02:49 lbr jdv79: episodical with DRAMA!
03:06 * kid51 must sleep
03:06 purl $kid51->sleep(8 * 3600);
03:14 wknight8111 joined #parrot
03:28 jjore joined #parrot
03:46 julian joined #parrot
03:51 julian joined #parrot
03:51 Coke joined #parrot
03:52 Coke bah. GC bugs in tcl again.
03:54 julian joined #parrot
03:58 * pmichaud svn up's to get a fresh copy of Parrot to take with him on the plane
04:00 julian joined #parrot
04:01 Coke plane back?
04:01 pmichaud yes
04:02 Coke safe travels
04:02 pmichaud thanks.  the next couple of days will be long stretches of nothing to do punctuated by rushing to get something done
04:03 julian joined #parrot
04:05 lbr pmichaud: svn up?  how do you manage without offline diff?
04:05 * lbr goes crazy
04:05 pmichaud I'm rarely offline.
04:06 lbr I mean svk has its many variances of suck, but at least you have a working copy :)
04:06 Coke I'm never offline long enough for it to matter.
04:06 pmichaud I have a working copy with svn, too.  But I suspect you mean "at least you can do intermediate commits."
04:06 lbr yes, but I do mean "svn diff"
04:07 pmichaud oh
04:07 pmichaud that's not actually something I typically need while writing
04:07 pmichaud probably just the way I program
04:08 * lbr uses it all the time on the modest assumption that he usually supplies the fuck ups not his team
04:08 * lbr is not always right
04:08 Coke svn diff works fine offline.
04:08 lbr really?
04:08 pmichaud coke:  for the most recent revision only, or even for historical ones?
04:09 Coke "local changes vs. whatever the last updated version"
04:09 pmichaud right
04:09 lbr which is the interesting bit
04:10 lbr if only svn could merge a liiiitle smarte
04:10 lbr r
04:10 pmichaud hmmm, it looks to me I could get a lot of spectests passing just by making them actually follow the spec :-)
04:11 Coke punk.
04:12 Ademan joined #parrot
04:20 julian joined #parrot
04:23 julian joined #parrot
04:28 julian joined #parrot
04:29 Coke seen tene?
04:29 purl tene was last seen on #parrot 1 day and 17 hours ago, saying: Oh, right, rakudo can't do IO yet.  [Feb 23 03:08:42 2008]
04:30 julian joined #parrot
04:34 Andy evening all
04:34 purl morning Andy
04:35 pmichaud morning Andy
04:35 Coke Hurm. I have a pmc that isa RPA. I also have a PIR sort method. I should be able to convince the pmc to sort itself without having to roll my own quicksort.
04:35 Andy see perlbuzz for stuff that people said about you
04:35 pmichaud Coke:  presuming that the elements of the RPA all know how to compare themselves?
04:36 Coke pmichaud: that's what the PIR sort method is for.
04:36 pmichaud oh, so passing the PIR sort method
04:36 Coke (I may want to compare them as integers, or alpha, or...)
04:36 Coke Right. looks like RPA inherits something from FPA that takes a c func.
04:36 julian joined #parrot
04:37 pmichaud Andy: I saw the perlbuzz article, many thanks!
04:38 Andy and I updated it with some other stuff
04:40 pmichaud I saw another posting somewhere where the person (a PHP guy) was disappointed that I didn't show "Hello world" actually running
04:41 pmichaud of course, my talk was already running a bit long, so I figured information was more useful than a "hello world" demo :-)
04:44 Coke [oops; continuation 0x3a20c00 of type 23 is trying to jump from runloop 530 to runloop 1]
04:44 Coke Someone remind me how to proceed here? =-)
04:45 pmichaud I never figured out how to resolve it.  The problem has something to do with :init :load, iirc
04:56 Coke Nothing like writing a test to say "this is how I want it to work" and having it work already.
04:57 Coke Now to figure out why the test works but the real code don't.
05:00 Coke does anyone else wish that IMCC gave a syntax for calling something in another namespace directly?
05:00 x joined #parrot
05:01 Coke (I find myself having to do a lot of get_root_global [], '' and then invoking it.
05:02 Andy ok, string_make_direct must die.
05:06 Tene Coke: pong?
05:06 Coke hio
05:17 jjore joined #parrot
05:28 * Coke 's eyes cross staring at tcl guts.
05:28 Coke ~~
06:55 stephen joined #parrot
06:55 stephen left #parrot
06:59 uniejo joined #parrot
07:04 uniejo joined #parrot
07:36 arcady I've finally beaten PGE/PCT into submission and gotten it to parse python tuples properly
07:42 arcady i should probably submit a patch for this
07:43 svnbotl r26054 | fperrad++ | trunk:
07:43 svnbotl : [Lua]
07:43 svnbotl : - add doc
07:43 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26054
07:44 arcady I can also without too much difficulty add PyList and PyTuple PMCs
07:49 UltraDM joined #parrot
07:49 svnbotl r26055 | fperrad++ | trunk:
07:49 svnbotl : [emacs]
07:49 svnbotl : - update keywords
07:49 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26055
07:57 IllvilJa joined #parrot
08:09 iblechbot joined #parrot
09:16 ilbot2 joined #parrot
09:22 ruoso joined #parrot
09:59 AndyA joined #parrot
11:06 AndyA joined #parrot
11:07 iblechbot joined #parrot
11:44 kj joined #parrot
13:24 AndyA joined #parrot
14:06 gryphon joined #parrot
14:06 Coke joined #parrot
14:06 mncharity joined #parrot
14:06 integral joined #parrot
14:06 Maddingue joined #parrot
14:06 nopaste joined #parrot
14:06 marmic joined #parrot
14:06 dwave_ joined #parrot
14:06 c9s_ joined #parrot
14:06 klapperl_ joined #parrot
14:06 jq joined #parrot
14:06 dalek joined #parrot
14:06 cognominal_ joined #parrot
14:06 skv_ joined #parrot
14:06 mj41_ joined #parrot
14:06 bgeron_ joined #parrot
14:06 Alias_ joined #parrot
14:06 uniejo joined #parrot
14:06 jjore joined #parrot
14:06 lbr joined #parrot
14:06 diakopter joined #parrot
14:06 amoore joined #parrot
14:06 buildbot joined #parrot
14:06 mdiep joined #parrot
14:06 svnbotl joined #parrot
14:06 cotto joined #parrot
14:06 teknomunk joined #parrot
14:06 cotto_ joined #parrot
14:06 particl1 joined #parrot
14:06 silug joined #parrot
14:06 Dave joined #parrot
14:06 avar joined #parrot
14:06 GeJ_ joined #parrot
14:06 cout joined #parrot
14:06 nnunley joined #parrot
14:06 bphillips joined #parrot
14:06 ewilhelm joined #parrot
14:06 jdv79 joined #parrot
14:06 arcady joined #parrot
14:06 kismet joined #parrot
14:06 peepsalot joined #parrot
14:06 Crassworm joined #parrot
14:06 shamu joined #parrot
14:06 TimToady joined #parrot
14:06 lidden joined #parrot
14:06 rblackwe joined #parrot
14:06 rhr joined #parrot
14:06 slavorg joined #parrot
14:06 cognominal joined #parrot
14:06 clunker joined #parrot
14:06 Khisanth joined #parrot
14:06 confound joined #parrot
14:06 spinclad joined #parrot
14:06 purl joined #parrot
14:06 workbench joined #parrot
14:06 leo joined #parrot
14:06 Piper joined #parrot
14:06 pmichaud joined #parrot
14:06 PerlJam joined #parrot
14:06 jonathan joined #parrot
14:06 jrockway joined #parrot
14:06 shorten joined #parrot
14:06 szbalint joined #parrot
14:06 pjcj joined #parrot
14:06 MagNET joined #parrot
14:06 TonyC joined #parrot
14:06 Juerd joined #parrot
14:06 wolverian joined #parrot
14:06 dngor joined #parrot
14:06 cj joined #parrot
14:06 cxreg joined #parrot
14:06 pfig joined #parrot
14:06 drforr_ joined #parrot
14:06 zev joined #parrot
14:06 zostay joined #parrot
14:06 tewk joined #parrot
14:06 BitPoet joined #parrot
14:06 Tene joined #parrot
14:06 rjbs joined #parrot
14:06 Sartak joined #parrot
15:11 Andy joined #parrot
15:12 Tene mdiep: ping
15:20 svnbotl r26056 | particle++ | trunk:
15:20 svnbotl : [runtime] build yaml dumper pbc file during make all
15:20 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26056
15:24 ruoso joined #parrot
15:24 Andy morning all
15:24 purl morning, Andy
15:28 mdiep Tene: pong
15:31 ruoso joined #parrot
15:39 jonathan hi all
15:39 * jonathan has made it to Holland, after a great time in the Ukraine.
15:43 po_boy joined #parrot
15:45 cognominal_ salut jonathan
15:46 jonathan cognominal_: hi :-)
15:48 cognominal_ how was the weather in Ukraine?
15:48 particl1 jonathan: we're totally on the same page. i just migrated t/spec/S12-methods/indirect_notation.t
15:51 jonathan particle: I actually discovered this half-implemented on my laptop when I got to Ukraine...I'm still not sure I remember starting the patch.
15:51 svnbotl r26057 | jonathan++ | trunk:
15:51 svnbotl : [rakudo] Implement indirect method calls of forms $x.'foo'() and $x.$foo().
15:51 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26057
15:51 jonathan cognominal_: I got lucky and had, for Ukraine in Feb, great weather.
15:52 jonathan The worst day was Saturday, which was the conference, so I was indoors most of that.
15:52 jonathan particle: You got many OO tests done?
15:53 particle no, just starting for the day
15:53 jonathan Oh, cool.
15:53 jonathan This will be helpful, thanks.
15:54 particle well, your patch will be helpful, because now i don't have to implement what that tests :)
15:55 jonathan Got a whole load of ideas for stuff to hack on for Rakudo.
15:56 particle great. let me know what it is.
15:56 jonathan Did you see Zev's patch for letting us write bits of the runtime in Perl 6?
15:56 particle it's one of my 35 open windows atm
15:57 mncharity (write bits of the runtime in Perl 6)++
15:57 particle i'd like to apply it with a real function in there, rather than Dummy.pm
15:58 particle something like List.max() seems appropriate
15:58 particle or steal something from the kp6 runtime
15:59 jonathan I don't think it will work too well writing classes that are part in PIR, part in Perl 6.
15:59 jonathan In fact, I *know* that won't work.
16:00 particle well, we can steal a whole class from kp6
16:00 jonathan Yes, that we can do.
16:02 particle http://svn.pugscode.org/v6/v6-Kinda​Perl6/src/KindaPerl6/Runtime/Perl6/
16:02 shorten particle's url is at http://xrl.us/bgqec
16:02 particle oops: http://svn.pugscode.org/pugs/v6/v6-Kin​daPerl6/src/KindaPerl6/Runtime/Perl6/
16:02 shorten particle's url is at http://xrl.us/bgqee
16:02 particle how about Signature?
16:02 purl Signature is, like, invalid
16:03 jonathan We don't have various stuff it needs implemented yet. :-(
16:03 particle well, it's a target to shoot for. however, List.pm is a good start
16:03 jonathan ?? !! for one.
16:04 particle we don't have ?? !! ? are you sure?
16:04 particle that's easy to implement, it's in PAST.
16:04 jonathan OK.
16:04 jonathan Ah, bike ride time...
16:04 jonathan back in a bit...
16:10 svnbotl r26058 | particle++ | trunk:
16:10 svnbotl : [rakudo] remove deprecated compiler directives replaced by t/spec/fudge
16:10 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26058
16:23 particle tene, timtoady: ping - there's a user in #perl6 to kickban
16:26 TimToady already did
16:26 TimToady well, I just kicked, didn't ban
16:26 particle probably good enuff
16:26 TimToady if not, I'll drag out my bin of bans
16:27 Tene didn't look malicous, just broken.
16:27 particle agreed
16:27 TimToady probably some harvester bot though
16:27 TimToady if so, they'll get a lot of grief from other channels too
16:28 particle the behavior wasn't malicious, the intent may have been
16:28 TimToady so still probably won't return, except as a lurker
16:28 TimToady presuming they fix the bug
16:29 kismet what user?
16:29 TimToady alc
16:29 kismet if you suspect it's a harvester, or publisher bot, and don't want it, just let me know, I'll get it glined.  Those are removed by request.
16:30 kismet and I need, to use, a few less, commas
16:30 TimToady well, the name was jobsearc
16:30 TimToady which is a bit odd
16:30 kismet muh-god I've become James Kirk.
16:32 Tene jonathan: I'm curious about why you used $S1000 instead of %r in your inline find_method call in your latest indirect method patch.
16:32 Tene ... oh, nevermind, I just need to read.
16:35 particle kismet: thanks, but that user/chan is on freenode
16:36 particle ...not sure if you're an op there.
16:36 sjansen joined #parrot
16:38 kismet nope.
16:38 kismet but I didn't see a reference to freenode either, so I assumed it was for here.  My fault for assuming.
16:38 particle no, harm, done.
16:39 TimToady well, #perl6 implies freenode, but not very hard :)
16:39 iblechbot joined #parrot
17:11 gryphon joined #parrot
17:24 grim_fandango joined #parrot
17:32 * Coke gets go ahead from his boss for the time off in yapc::na. (good thing as the flight is booked.)
17:34 Coke and an indirect nod on yapc::eu.
17:38 kj joined #parrot
17:42 Theory joined #parrot
17:49 davidfetter joined #parrot
17:49 davidfetter re
17:49 davidfetter anybody in lisbon atm?
17:50 davidfetter anybody else*
17:54 Coke .pt or .ct.us ?
17:55 davidfetter .pt
17:55 * davidfetter managed to miss all the parrots @ FOSDEM :P
17:55 Coke nope. (I could drive to .ct.us in a pinch.)
17:55 davidfetter heh
17:57 davidfetter are you in .ct?
17:57 * davidfetter used to live in madison...
17:58 Coke born and raised in .ct; currently living in .ny (upstate)
17:58 Coke You can probably see most of the places I've lived from the top of the hotel at the mohegan sun. =-)
17:58 davidfetter heh
17:59 * davidfetter has lived varied enough places that you can't really see them all at once from space
18:00 Coke ... what an excellent idea. I'm going to put in a few of the most geographically disparate ones, and see how high I have to go in google earth to see them all!
18:01 davidfetter heh
18:01 Coke does petergibbs show up on IRC?
18:01 davidfetter SELECT clue FROM davidfetter;
18:01 davidfetter ---
18:01 davidfetter 0 Rows
18:07 Coke seen kid51?
18:07 purl kid51 was last seen on #parrot 15 hours and 32 seconds ago, saying: must sleep
18:07 Coke opbots, names
18:10 teknomunk joined #parrot
18:16 mncharity Hi, could I get a reality check?  I'd like to have a script which takes a file with p6 grammar code (grammar Foo {...}), a grammar name (Foo), a start rule, and an input file, and runs the start rule with the input, yielding a Match tree.  I'd then like to dump the Match tree to stdout, ideally as yaml.  My impression of how to go about this is
18:18 mncharity use compilers/pge/ , and then... (hmm, ok, what i was thinking now seems invalid)... suggestions?
18:20 mncharity demo.pir mutant if it can (be made to)? accept a full grammar?
18:21 mncharity download nice CPAN module?
18:24 mncharity use rakudo?
18:29 davidfetter anybody seen flavio?
18:39 Tene mncharity: if we get file IO supported in rakudo, it should be a pretty short Perl 6 script.
18:39 Tene Or you could write it in PIR using PGE
18:40 cognominal_ the thing missing is all the pair syntaxes
18:40 cognominal_ but one can get away with  :key( 'value' )
18:45 mncharity Tene: do you have a feel for when IO might be supported?  say, {February, March, later}?
18:46 Tene mncharity: I haven't looked to see if Parrot supports it yet or not.  I was planning on working on it tonight or tomorrow night.
18:46 mncharity ooo, nifty.  ok, I'll wait then.
18:48 mncharity re 'you could write it in PIR using PGE', given, say, a grammar and input fed as single lines, does it seem like this would be a two-ish line change to demo.pir?  one to do grammar(s?) rather than a rule, and one to invoke the grammar rule?  or is it "more interesting" than that?
18:56 cognominal_ Tene, I don't see io specified in http://perlcabal.org/syn/S29.html
18:56 cognominal_ on the other hand I see some tests in pugs
18:58 cognominal_ ho, http://search.cpan.org/~lichtkind/Perl6-​Doc-0.35_1/lib/Perl6/Doc/Design/S16.pod
18:58 shorten cognominal_'s url is at http://xrl.us/bgqng
18:58 cognominal_ not on dev.perl.org :(
19:01 particle cog: http://perlcabal.org/syn/S16.html -- IO
19:01 particle mncharity: we don't have a way to dump a match tree in yaml yet
19:02 particle i'm working on something that'll dump AST as yaml.
19:06 cognominal_ don't we have json?
19:06 mncharity re 'dump a match tree in yaml', ok.  I could always post-process another format into yaml.
19:07 particle we've decided to use a subset of yaml that'll look much better than json
19:07 particle it's basically json+
19:07 mncharity re 'dump AST as yaml', yay.  That's a separate objective, but perhaps the code can be applied to Match's as well.
19:07 particle yes, the yaml dumper is generic
19:08 mncharity :)
19:08 Ademan joined #parrot
19:09 cognominal_ json does not handle cicurlarity while yamld does if I recall correctly
19:09 particle yes, however i haven't addressed circularity yet
19:09 particle json doesn't address typing well
19:10 particle with yaml i can say: !PAST::Op { ... }
19:10 Tene docs/ops/io.pod looks good.
19:11 particle also json requires strings to be quoted, which is ugly imo and unnecessarially verbose
19:11 Tene Should be pretty straightforward to add some basic IO functions to rakudo
19:11 particle it'd be great to have spectests available before having the implementation
19:12 Tene Yes, it would.
19:12 Tene mncharity should do it.
19:12 cognominal_ some are in pugs
19:12 particle looks like there's t/spec/S16-* already
19:12 cognominal_ yes
19:13 particle however, some smartlinked tests here (http://perlcabal.org/syn/S16.html) are not in t/spec/ yet
19:13 cognominal_ problem is that rakudo does not have pairs yet
19:13 particle search for "Show t/"
19:14 Tene If there are tests for 'open', 'read', and friends by tonight, I'll spend an hour or two working on implementing them tonight.
19:14 cognominal_ these tests use them intensively
19:14 * particle wonders if kp6 has Pair.pm
19:14 particle http://svn.pugscode.org/pugs/v6/v6-KindaP​erl6/src/KindaPerl6/Runtime/Perl6/Pair.pm
19:14 particle yay!
19:14 shorten particle's url is at http://xrl.us/bgqn5
19:14 cognominal_ we need to parse them too
19:14 barney joined #parrot
19:14 Tene cognominal_: they're not in spec/S16-* yet.
19:14 particle let's see if i can hack zev's patch to include Pair.pm
19:15 particle yes, we need to parse them too, but they're in STD
19:15 Tene What do we need pairs for specifically right now?
19:15 cognominal_ I would be curious how you do it, I tried to no avail...
19:16 * barney needs Pairs for Eclectus
19:17 cognominal_ in French, paire means balls too :)
19:17 cognominal_ time for particle to prove he have some.
19:17 Tene I'm curious about what 'method true' in that Pair.pm is for.
19:17 cognominal_ s/have/has/
19:18 particle .true returns Bool::True in Pair
19:18 Tene What... what is that good for?
19:18 particle defined testing
19:18 Tene Oh.
19:19 cognominal_ btw, I have token noun {
19:19 cognominal_ 'True'  {*}                                #= True
19:19 cognominal_ || 'False' {*}                                #= False
19:19 cognominal_ || <colonpair> {*}                            #= colonpair
19:19 cognominal_ #   | <fatarrow>  {*}                            #= fatarrow
19:19 cognominal_ but I don't know what to do for boolean actions
19:20 cognominal_ I find strange that booleans litterals starts by a capital because it is reserved for namespaces
19:20 lichtkind joined #parrot
19:21 lichtkind pmichaud: cheers how is doing rakudo?
19:21 particle you should whack TimToady over the head with that q.
19:21 particle that was for cog, not lichtkind
19:21 lichtkind allrigth :)
19:22 particle as for boolean actions for True|False...
19:22 TimToady true() is a function
19:22 cognominal_ so what is True?
19:22 TimToady True is an enum, which is both a type and a value
19:22 particle ...Bool class sets up globals for True and False
19:23 TimToady enums are primarily considered singleton types
19:23 particle so, my True $x; # automatically contains True?
19:23 TimToady in the sense of representing only one value
19:23 TimToady yes
19:23 particle don't need to say my Bool::True $x
19:23 particle ok
19:23 lichtkind pmichaud: where are lieing the userdocs on parrot you told me about
19:23 particle er, i guess that'd be my Bool $x = True
19:23 TimToady and in signatures, values are even considered types in that sense
19:24 TimToady so you can have a parameter of "42" that has to match
19:24 TimToady which one uses all the time in recursive multis for the end condition
19:24 cognominal_ each time I hit TimToady with an apparent inconstency, he hits me back with a deeper consistency :)
19:24 TimToady we've been thinking about this a while now :)
19:25 TimToady it's not entirely consistent yet, but it can never be completely consistent on all levels
19:25 particle lisp begs to differ
19:25 TimToady lisp only has one level
19:25 ewilhelm and no syntax
19:25 cognominal_ but when you do  my $x = True,   you get a type or a value?
19:26 TimToady yes
19:26 particle now you're just having fun :)
19:26 TimToady It's mainly a value pretending to be a type
19:26 TimToady or maybe it's the other way around...
19:27 ewilhelm it's not just a typed value?
19:27 cognominal_ depends on the way you use $x?
19:27 particle $x isn't constrained only to hold True, though.
19:27 particle with "my True $x;" it is
19:27 TimToady yes, but that's a container type, not a value type
19:27 particle it's effectively constant
19:27 particle right.
19:28 TimToady there's it's using True as a real subset type
19:28 particle in "my $x = True" you get the container typed because it's value is.
19:28 TimToady subset True is Bool where {.true} or some such
19:28 peeps[work] joined #parrot
19:29 TimToady subset February of Month where 2;
19:29 TimToady an enum is just a list of those
19:29 cognominal_ it was just a matter of time that in scripting lanuguage everytin become a role. The only thing hopefully missing is the casting.
19:30 TimToady a type that is constrained to a single value may be used as that value
19:30 TimToady and protoobjects are constrained to be undefined
19:30 particle perl 6 breaks the mold, so there's no casting :P
19:31 TimToady the only question is whether its toxic mold...
19:31 TimToady *it's
19:32 particle we'll know if programmers come down with sick language syndrome
19:32 particle Use of uninitialized value in hash element at c:/usr/local/parrot/clean/l​anguages/perl6/t/spec/fudge line 184, <> line 117.
19:32 particle hrmm.
19:33 particle how can i have too many windows and not enough windows open at the same time?
19:33 x joined #parrot
19:33 cognominal_ from the dict but that could be Dilbert  " take the marzipan and mold it into a cone shape."
19:37 * particle wants a test harness that works better for fudge results
19:38 * particle wished *nix supported negative exit status codes
19:40 grim_fandango joined #parrot
19:41 grim_fandango joined #parrot
19:52 parrot-poke joined #parrot
19:53 particle cognominal_: wrt 'True' action, it'd be something like 'self'
19:53 particle the Bool class creates globals
19:54 particle you can do inline pir to '%r = get_global ...'
19:59 cognominal_ in fact it should not be a noun
19:59 cognominal_ it must be in the hndling of typename
20:00 particle neither should 'self' or 'undef' be a noun
20:03 cognominal_ hum, in the last svn checkout  I don;t get a prompt when starting rakudo :(
20:03 svnbotl r26059 | bernhard++ | trunk:
20:03 svnbotl : [Plumhead]
20:03 svnbotl : Remove abandoned variant 'yacc'
20:03 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26059
20:03 cognominal_ I lost hope to get parrot working with svk
20:06 cognominal_ hum inlining pir in actions, still new to me
20:08 schmalbe joined #parrot
20:15 barney joined #parrot
20:35 arbingersys joined #parrot
20:44 mire joined #parrot
20:53 svnbotl r26060 | bernhard++ | trunk:
20:53 svnbotl : #51056: [PATCH] coding style consistency and doc fix
20:53 svnbotl : Courtesy of Christoph Otto.
20:53 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26060
21:00 jonathan particle: self is a noun for now because it should be a term, but a term should be a noun, and if you try and do that you get the parser into some infinite recursion. Or something. There's a comment in the grammar about this.
21:01 particle sure, my point was that we cheat. working trumps good.
21:03 particle zev's patch isn't working as-is
21:05 jonathan Ah?
21:05 jonathan :-S
21:05 particle a few small problems
21:05 particle system.pir doesn't exist, np, just remove it from the patch
21:06 jonathan Ah, OK.
21:06 particle but 'echo "" > foo.pir' is actually putting "" in the file
21:06 particle when it should be blank
21:06 particle so, i need to find a portable way to create an empty file
21:06 particle looking at ExtUtils::Command now
21:07 particle hoping 'touch' will do it
21:07 kj particle: what if you create a pir file with a comment in it?
21:07 kj would that help?
21:07 jonathan particle: touch won't work on win32?
21:07 particle ExtUtils::Command's touch will work, with perl
21:07 particle win32 doesn't have a 'touch' command in the shell
21:08 particle kj: trying now
21:09 particle kj: no, echo still displays the quotes
21:10 particle no quotes seems to work on linux and win32
21:13 ewilhelm particle: that's a file with "\n" in it
21:13 particle ?
21:13 ewilhelm echo > file; wc file
21:13 ewilhelm has one character -- not empty
21:13 particle echo "# foo bar" > file ## displays "\"# foo bar\"" on windows
21:14 particle the quotes end up inside the file
21:14 kj what if you also echo a backspace char :-)
21:14 ewilhelm did you want an empty file or a file with no quotes in it?
21:14 ewilhelm try `>file`
21:14 particle echo # foo bar > file ## displays "# foo bar" on windows and linux
21:14 ewilhelm yeah, but that's not empty
21:15 particle doesn't need to be empty
21:15 particle just needs to be a portable command
21:15 particle something that's parsable pir
21:15 ewilhelm ok, thought you started with `touch`
21:15 particle empty, or a comment, or noop
21:15 * ewilhelm $lurk = 1
21:15 particle yes, because an empty file is parsable pir. sorry that wasn't clean
21:16 particle s/n$/r/
21:18 kj particle: a few days ago you mentioned auto generating nci stubs, in relation to a c parser. I didn'tand don't quite understand what you meant
21:19 kj (as I now have c99 running, I might as well see what I can do on the nci stub stuff)
21:20 particle ok, see for example runtime/parrot/library/SDL.pir
21:21 particle dlfunc sdl_function, libstl, ...
21:22 particle if we could parse the header files and generate those lines, it'd make creating nci libraries much easier
21:22 Tene kj: pdds/draft/pdd16_native_call.pod
21:22 Tene Look at the SD... yeah, what particle said.
21:23 particle also if we could generate things like _set_Color_layout
21:23 particle the structs defined in the c headers
21:24 kj mm.ok what header files are you referring to?
21:24 kj so, I fail to see what needs to be parsed, and what needs to be generated
21:24 particle grab the libSDL package and look at SDL.h
21:30 kj i guess you mean the sdl code from sdl.org
21:31 particle rpm -qa | grep SDL
21:31 Tene Looks like kj is on windows
21:31 kj yes he is ;-)
21:31 particle ah, right. windows.
21:31 Tene That's always disorienting to me.
21:31 particle libsdl.org
21:32 kj yeah got it
21:32 Tene http://pleasedieinafire.net/~tene/SDL.h
21:32 kj so, for instance, consider:
21:32 kj extern DECLSPEC void SDLCALL SDL_QuitSubSystem(Uint32 flags);
21:32 kj from sdl.h
21:33 kj what kind of PIR is needed for this example?
21:35 kj ah i see (scrollback; missed that)
21:36 Tene Look at... SDL_video.h
21:36 Tene look for extern DECLSPEC SDL_Surface * SDLCALL SDL_SetVideoMode (int width, int height, int bpp, Uint32 flags);
21:36 Tene That matches up with:
21:36 Tene dlfunc sdl_function, libsdl, 'SDL_SetVideoMode', 'piiil'
21:36 Tene in SDL.pir
21:37 Tene the first character in the last item is the return type
21:37 Tene pointer
21:37 Tene then the types of the arguments
21:37 Tene int, int, int, long
21:37 kj tene: i see.thanks
21:38 kj is that really a pain to have to write manually?
21:38 Tene [tene@kupo ~]$ find /usr/include/ -name '*.h' | wc -l
21:38 Tene 4476
21:40 kj yeah, but obviously you're not going to invoke ALL functions on your system.
21:40 kj are you?
21:40 kj :-)
21:40 Tene Simultaneously.
21:40 purl simultaneously is pleased and scared
21:58 svnbotl r26061 | kjs++ | trunk:
21:58 svnbotl : [c99] fix grammar rule to parse structs
21:58 svnbotl : have # directives skipped for now through whitespace rule. not sure how to handle preprocessor; inline or as a separate parser.
21:58 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26061
22:20 Coke (generating an empty file) we already do this in pge, I think.
22:21 Coke hurm. I see TOUCH in the makefile, but it's not where I thought it as.
22:31 Limbic_Region joined #parrot
22:32 Coke Limbic_Region: hey, Joshua. Someone was looking for you in the past week.
22:32 Coke ... ooh, we have logs you can scan. =-)
22:40 IllvilJa joined #parrot
22:42 Limbic_Region any idea who it might have been?
22:42 Limbic_Region I got an email from robert concerning piper
22:42 Limbic_Region haven't had a chance to provide a patch
22:43 Limbic_Region other than that, no idea
22:43 Limbic_Region they didn't leave a message for me via purl
22:53 Limbic_Region Coke - where is the search functionality?  I can only seem to find the raw logs
23:01 * Limbic_Region can't be bothered to wait and decides to write some WWW::Mechanize instead
23:07 wknight8111 joined #parrot
23:12 Limbic_Region Coke - found it
23:13 Limbic_Region Coke - it was jmcadams on 2008-02-22
23:13 Limbic_Region "<jmcadams> Conrad Schneiker or Joshua Gatcomb hang out here?"
23:29 Limbic_Region so are the logs searchable or not?
23:45 iblechbot joined #parrot
23:47 kid51 joined #parrot
23:53 cj joined #parrot
23:55 mire joined #parrot
23:56 svnbotl r26062 | jonathan++ | trunk:
23:56 svnbotl : [rakudo] Make $.x used away from declaration with has do something closer to the right thing.
23:56 svnbotl diff: http://parrotvm.org/svn/parrot/revision/?rev=26062

| Channels | #parrot index | Today | | Search | Google Search | Plain-Text | summary

Parrot | source cross referenced