Camelia, the Perl 6 bug

IRC log for #parrot, 2008-08-18

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:09 AndyA joined #parrot
00:12 kid51 purl nopaste
00:12 purl it has been said that nopaste is at http://nopaste.snit.ch/ (ask TonyC for new channels) or http://rafb.net/paste or http://poundperl.pastebin.com/ or http://paste.scsys.co.uk/ or App::Nopaste or tools/dev/nopaste.pl
00:17 nopaste "kid51" at 71.247.53.241 pasted "Parrot::Configure::Trace::trace_data_c() applied to 'linkflags'" (78 lines) at http://nopaste.snit.ch/13823
00:18 kid51 Note:  For > 1 year, I absolutely had to build with this script to avoid problems like the editline vs readline one.
00:18 tetragon kid51: My system isn't contaminated by either fink or darwinports, so none of the checks that look there can add those locations
00:18 tetragon And the GMP, readline, et al checks don't use integrate
00:19 kid51 In recent months I have been able to build without the extra flags, but I get cleaner 'make' output with them.
00:19 tetragon What's manicheck
00:20 kid51 nomanicheck:  Skip the very first step:  verifying MANIFEST.  Slow on old Mac.
00:21 tetragon And since you didn't set linkflags at the command line, your trace wouldn't show any change at inter::progs
00:22 tetragon Although, I think you would see a change if you checked the value of 'link'
00:23 bacek joined #parrot
00:23 kid51 I wonder if it's really a good idea for the darwin hints file to cause suppression of:  -L/opt/local/lib
00:24 kid51 Can you provide a definition of:  '-undefined dynamic_lookup'?
00:25 tetragon It's in the ld manpage
00:25 tetragon http://developer.apple.com/documentation/​Darwin/Reference/ManPages/man1/ld.1.html
00:26 tetragon What dynamic_lookup does is tell ld/gcc not to give an error if it can't find a symbol at linktime.  Instead, you get the error at runtime if the symbol still isn't found
00:26 kid51 What does --ld do different from --link ?
00:27 kid51 What is definition of:  linktime  runtime   -- in this context?
00:27 tetragon Runtime is after the build is finished and I'm running a binary
00:27 tetragon linktime is during the build
00:28 tetragon That was one of the issues behind http://rt.perl.org/rt3/Tic​ket/Display.html?id=52212
00:28 NotFound --ld is used to build the library, --link to build the executables, if I remember well.
00:29 tetragon (however, the readline check does not appear to function the way it did when that report was created)
00:30 kid51 So, in this context, since "linktime is during the build," does that mean that it occurs during 'make'?
00:36 dalek r30289 | jkeenan++ | trunk:
00:36 dalek : Correct one misleading statement in inline comments:  %flags --> .
00:36 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30289
00:36 tetragon Yes (or Configure.pl)
00:37 kid51 Well, to answer one of my earlier questions:  config/init/hints/darwin.pm explicitly sets a value for linkflags:          linkflags           => '-undefined dynamic_lookup',
00:38 kid51 Thereby overwriting anything captured from the Perl 5 defaults.  And since we probably went through a lot of sweat to deduce this value, I think we should continue to leave it alone.
01:08 rurban_ joined #parrot
01:09 Ademan joined #parrot
01:48 Zaba joined #parrot
02:03 Zaba joined #parrot
02:06 Auzon joined #parrot
02:13 bacek joined #parrot
02:15 dalek r30290 | rgrjr++ | trunk:
02:15 dalek : [comment]  Change exception_handler to ExceptionHandler in
02:15 dalek : instruction_writes.
02:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30290
02:16 dalek r30291 | rgrjr++ | trunk:
02:16 dalek : [docs]  Change exception_handler to ExceptionHandler.
02:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30291
02:19 dalek r30292 | rgrjr++ | trunk:
02:19 dalek : [docs] Two minor tweaks:
02:19 dalek :    + Explain that "make codingstd_tests" may fail, as long as none of
02:19 dalek :      the "make test" subset do.  chromatic++, kid51++ for feedback.
02:19 dalek :    + Fix "Sourc0" typo.
02:19 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30292
02:40 Zaba joined #parrot
02:44 tetragon joined #parrot
03:13 tetragon joined #parrot
03:14 Zaba_ joined #parrot
04:00 Zaba joined #parrot
04:35 Theory joined #parrot
05:18 jkva joined #parrot
05:25 tetragon joined #parrot
05:32 ank joined #parrot
05:36 Psyche^ joined #parrot
05:49 Zaba joined #parrot
05:58 uniejo joined #parrot
07:06 masak joined #parrot
07:13 tetragon joined #parrot
07:23 Ademan joined #parrot
07:27 ank joined #parrot
07:59 dalek St�phane Payrard | dtrace:
07:59 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
08:47 Zaba joined #parrot
08:48 dalek St�phane Payrard | dtrace:
08:48 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
08:51 dalek St�phane Payrard | dtrace:
08:51 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
08:51 Zaba_ joined #parrot
08:52 dalek St�phane Payrard | dtrace:
08:52 dalek link: http://www.perlfoundation.o​rg/parrot/index.cgi?dtrace
08:57 Zaba joined #parrot
08:58 Ademan joined #parrot
09:04 paco joined #parrot
09:09 rurban_ joined #parrot
09:33 Zaba joined #parrot
09:37 dalek r30293 | kjs++ | trunk:
09:37 dalek : [pirc] many updates
09:37 dalek : + update README.pod
09:37 dalek : + many POD comments added
09:37 dalek : + many updates in parser-back-end communication
09:37 dalek : + handle .const directive
09:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30293
09:45 Whiteknight joined #parrot
09:48 gaz joined #parrot
09:57 barney joined #parrot
10:26 dalek r30294 | pmichaud++ | trunk:
10:26 dalek : [rakudo]:  Fix bug with list context in statement modifier, Array.values
10:26 dalek : * Resolves RT #57960 (masak++)
10:26 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30294
10:30 Zaba joined #parrot
10:36 Zaba_ joined #parrot
10:52 Zaba joined #parrot
10:56 Whiteknight joined #parrot
11:00 Zaba_ joined #parrot
11:19 donaldh joined #parrot
11:21 Zaba joined #parrot
11:50 dalek r30295 | pmichaud++ | trunk:
11:50 dalek : [rakudo]: spectest-progress.csv update: 121 files, 2220 passing tests
11:50 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30295
11:50 * jonathan waves at pmichaud
11:50 jonathan Good journey back?
11:52 pmichaud yes, not too bad
11:52 pmichaud I just about have precompiled Test.pbc working :-)
11:53 pmichaud managed to get some very good hack time on the plane
11:53 jonathan Nice!
11:53 jonathan pmichaud++
11:53 pmichaud barring any major interruptions around the house (none expected), I'll have it committed + running today.
11:55 jonathan yay!
11:55 pmichaud how was your journey?
11:55 jonathan I'm pondering Rakudo-daying on Wednesday. Does that work for you?
11:55 jonathan Mine was fine.
11:55 pmichaud Wednesday works *very* well for me
11:56 jonathan Flight was short and left on time.
11:56 jonathan The train connections back here worked out just nicely.
11:56 jonathan The right amount of time to find a snack in Vienna, but not so long I was waiting around loads.
11:56 pmichaud at the moment my sleep schedule has me still getting up around 4am (even though I'm now -7 hours from Copenhagen)
11:56 rurban_ joined #parrot
11:57 pmichaud this morning I woke up at 2am (?!?)
11:57 jonathan ?!
11:57 jonathan Wow!
11:58 pmichaud anyway, Wednesday sounds very good.  Also because that comes immediately following this month's release, so we don't have to be quite so careful about commits :-)
11:58 pmichaud I need to update NEWS for Rakudo
11:59 pmichaud and PCT
12:00 AndyA joined #parrot
12:01 jonathan Yay, I can Break Stuff!
12:02 pmichaud I CAN HAZ FAIL?
12:02 jonathan YOU CAN HAZ EPIC FAIL
12:02 jonathan Hey, I think we just designed the catch/throw syntax for LOLCODE exceptions!
12:03 pmichaud LOLCODE exception handling is (I think)   O NOES
12:03 jonathan Oh, it already has it?
12:03 jonathan Nice
12:03 pmichaud oh, out of curiosity -- how hard is it to get from where you are to Australia?  ;-)
12:04 jonathan I don't know that it's *hard*. I do know that it's long.
12:04 jonathan And probably not cheap.
12:04 pmichaud okay, that's what I was curious about.
12:05 pmichaud On the way home from the Airport Paula was talking about the various trips she'd like to take over the next 4-12 months... :-)
12:06 jonathan Generally, it looks like it costs �1000 - �1250 GBP range, 24 hours or so of journey and two connections.
12:07 pmichaud ouch
12:07 pmichaud okay, we'll go for something a bit less onerous for you :-)
12:07 jonathan A lot of them seem to go via London.
12:07 jonathan Which is suckful in that, that means going 2+ hours in the wrong direction.
12:08 pmichaud our kids are already asking about our next vacation trip, and Australia seems to be high on the list :-)
12:09 jonathan The sanest option appears to involve me going to Frankfurt and taking the flight from there.
12:10 jonathan That's still 21 hours from there.
12:10 pmichaud okay.  I'd give it a low probability of occurrence then, unless there's an Australia-based conference that is worth attending.
12:10 jonathan The plane makes one stop, but it's not a connection - you continue on the same plane.
12:10 jonathan Yeah
12:10 jonathan It's a real long trek out there for me.
12:10 pmichaud well, it's long from here also :-)
12:10 jonathan Sure. :-)
12:11 cognominal well, if you got access to a power plug. That means a lot of time coding
12:11 cognominal without much distractions
12:12 pmichaud I _think_ that from now on I'm going to try hard to get "economy plus" bookings on airplanes, or whatever would get me a power port
12:12 rurban_ joined #parrot
12:12 Zaba joined #parrot
12:12 pmichaud it'd also be less likely that I'd get bumped from my preferred seat, which is what happened to me on the Chicago->Copenhagen flight :-|
12:15 dalek r30296 | kjs++ | trunk:
12:15 dalek : [NEWS] update some PIRC news items.
12:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30296
12:16 jonathan :-(
12:16 jonathan OK, need to go to the store and get some nom...the fridge is bare..
12:16 pmichaud yes, I need nom also -- but no-one else is awake here yet
12:16 dalek r30297 | pmichaud++ | trunk:
12:16 dalek : [pct]:  Correct order-of-sub generation for POST.
12:16 dalek : This update causes POST to generate Parrot subs in the
12:16 dalek : order in which they appear in the tree (previously sibling
12:17 dalek : subs would occur in reverse sequence in the output, which
12:17 dalek : could break :load/:init ordering).
12:17 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30297
12:17 Tene Argh, .HLL-map timing issues. :(
12:17 kj joined #parrot
12:18 dalek r30298 | pmichaud++ | trunk:
12:18 dalek : [pct]:  Add 'loadinit' attribute and code generation to PAST::Block nodes
12:18 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30298
12:25 rurban_ joined #parrot
12:38 dalek r30299 | bernhard++ | trunk:
12:38 dalek : [Pipp] Start with support for while- and for-loop
12:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30299
12:43 dalek r30300 | julianalbo++ | trunk:
12:43 dalek : refactor PDB_script_file
12:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30300
12:49 dalek r30301 | bernhard++ | trunk:
12:49 dalek : [doc] Add news about Pipp.
12:49 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30301
12:53 Zaba joined #parrot
12:57 Zaba_ joined #parrot
13:19 desertmax joined #parrot
13:22 dalek r30302 | coke++ | trunk:
13:22 dalek : [docs] - Reorder sections. Core implementation is very important; Core deprecation is least important. Update Tcl section. Mention removal of implicit method dispatch; Reduce importance of new say opcodes.
13:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30302
13:25 dalek r30303 | coke++ | trunk:
13:25 dalek : [docs] ruthlessly cut my two line tcl bullet into one line.
13:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30303
13:31 Zaba joined #parrot
13:39 AndyA joined #parrot
13:40 smash joined #parrot
13:41 smash hello everyone
13:43 Topic for #parrotis now http://www.parrotcode.org | Last Release: 0.6.4 "St. Vincent Amazon"
13:43 Zaba joined #parrot
13:47 Zaba_ joined #parrot
13:50 pmichaud hello, smash
13:51 moritz hi
13:51 smash pmichaud: so, back home already ?
13:52 pmichaud yes -- got back in yesterday afternoon
13:53 smash nice
13:53 smash i got back yesterday at 6am
13:53 moritz has YAPC::EU been productive?
13:54 dalek r30304 | julianalbo++ | trunk:
13:54 dalek : fix perl6multisub.pmc for c++ builds
13:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30304
13:55 smash moritz: a lot, at least for me
13:57 moritz very good
13:59 gryphon joined #parrot
14:02 skv joined #parrot
14:02 jonathan julianalbo++ # fixing Perl6MultiSub for C++ without breaking it, now I added tests
14:06 pmichaud yapc::eu was very productive for me
14:06 pmichaud not much code (yet), but lots of good design and other fun
14:07 pmichaud -Ofun remains important :-)
14:07 jonathan As does --target=lolcode ;-)
14:08 Tene So, there's a problem with .HLL_map timing, maybe?
14:08 pmichaud I don't understand ".HLL_map timing"
14:08 Tene HLL_map takes place at compile-time, but most languages right now are creating classes a truntime.
14:08 pmichaud oh
14:08 Tene So .HLL_map doesn't work unless you set the sub that creates the classes to :immediate
14:08 pmichaud yes, there's probably a problem there.  I can see two options
14:09 Tene I was trying to get that working with cardinal on the plane last night, but there are issues.
14:09 pmichaud (1) allow HLL mapping to be set at runtime
14:09 pmichaud (2) have HLL mapping tie to a namespace so it can be deferred to the point when the class is created (similar to what we do now for :method)
14:10 Tene Is there ever going to be a time when we might care about setting hll mapping at runtime, do you think?
14:10 pmichaud being able to introspect/modify HLL mapping at runtime might be useful/important in any case
14:10 pmichaud introspecting HLL mapping is definitely important.  compiler tools and utilities might need to be HLL aware
14:11 pmichaud and it would seem to be easier to adjust HLL mapping at runtime than to try to update IMCC to do it statically at compile time
14:12 pmichaud I think it's worth a discussion on the m/l, and probably a ticket
14:15 Tene Okay, I'll try to get that started today.
14:17 mmcleric left #parrot
14:18 pmichaud P6object could even provide a 'hllmap' option
14:18 pmichaud so that we add the hll mapping when the class is created
14:18 pmichaud (and the mechanics are abstracted)
14:26 Tene That's a very nice option.
14:27 pmichaud if you don't draft a message for it later, I probably will
14:27 pmichaud it's going to be important for rakudo soon
14:33 nopaste "pmichaud" at 76.183.97.54 pasted "anyone know what this produces under windows?" (4 lines) at http://nopaste.snit.ch/13825
14:34 smash pmichaud: want me to fire up the windows virtual machine and test it ?
14:34 pmichaud that would be helpful, yes
14:34 pmichaud I just need some idea of if there's any fractional part in the number (there is under linux)
14:34 smash ok, give me a couple of minutes
14:35 * smash nods.
14:36 jonathan pmichaud: IIRC, it doesn't produce a fractional part...also it returns on Windows with a different epoch, unless rblasch got to fixing that already.
14:38 Zaba joined #parrot
14:45 pmichaud I'm not worried about the epoch -- just trying to figure out how I can generate unique identifiers.
14:46 pmichaud in the sense of "universally unique"
14:46 smash pmichaud: 1.28635e+10
14:46 pmichaud oh, well that's terribly unhelpful.
14:46 smash but, wait.. let me update to last revision
14:46 pmichaud under linux I get "1219069980.072166
14:46 pmichaud "
14:47 smash yeap, i get "1219070839.746350" in linux
14:47 smash (svn up, building, etc on windows)
14:48 pmichaud (knowing what Windows produces is _very_ helpful, the fact that it returns an e+10 number is very disappointing.)
14:49 jhorwitz joined #parrot
14:50 jonathan pmichaud: Once epoch is fixed to be 1970 rather than 1600, I expect it will be rather better. ;-)
14:50 smash my windows's parrot was a bit outdated, cross your fingers ;)
14:50 pmichaud jonathan: you think the extra digit is what caused e+10 ?
14:51 rdice joined #parrot
14:51 jonathan purl: 4000*365.25*24*3600
14:51 purl 126230400000
14:51 nopaste "pmichaud" at 76.183.97.54 pasted "floating precision on linux" (11 lines) at http://nopaste.snit.ch/13826
14:51 jonathan purl: 30*365.25*24*3600
14:51 purl 946728000
14:52 jonathan Rather different order of magnitude. :-)
14:52 pmichaud e+10 versus e+08, yes?
14:52 pmichaud and why 4000?
14:53 pmichaud shouldn't that just be 400?
14:53 jonathan Oh
14:53 jonathan Yes
14:53 jonathan :-)
14:53 jonathan purl: 400*365.25*24*3600
14:53 purl 12623040000
14:53 pmichaud but even on my box...
14:54 nopaste "pmichaud" at 76.183.97.54 pasted "floating precision on linux, #2" (11 lines) at http://nopaste.snit.ch/13827
14:54 pmichaud I'm a little surprised that the windows version of Parrot would choose to do e+10
14:55 smash hmm, getting errors now when running Configure.pl
14:55 smash something changed since the last time i builded parrot on windows
14:57 pmichaud maybe I need to get parrot running on windows here
14:57 pmichaud at any rate, I need something that I can use to generate unique sub names in Parrot
14:58 pmichaud (needed for .pbc files to work, fwiw)
14:58 smash alas, even with errors it builds :)
14:58 smash pmichaud: 1219073042.140000
14:59 smash looks someone fixed that e+10 thing
14:59 smash cool
14:59 pmichaud and subsecond times, too!
14:59 pmichaud yay!
14:59 pmichaud only to .01 second, but that's probably good enough.
15:01 smash pmichaud: i'll leave the windows machine powered on, just ping if you need anything else
15:01 jonathan Probably will do for now. :-)
15:11 Zaba joined #parrot
15:12 jhorwitz pmichaud: out of curiousity, what do you need unique sub names for (sorry if i missed something)
15:13 particle generated non-anonymous subs must have unique names
15:14 particle within the same namespace
15:14 particle currently, every pct-generated file starts with a sub _block11
15:15 particle if these files are .included...
15:15 particle you get an earth-shattering kaboom.
15:19 * jhorwitz ducks and covers
15:19 * jonathan observes the earth is still not shattered and figures everyone had the sense not to try this yet
15:20 jhorwitz is the LHC running yet?  ;-)
15:20 pmichaud jhorwitz: did jonathan's explanation make sense or do you want my version?
15:20 jhorwitz particle's explanation makes sense.  :)
15:20 pmichaud oh, sorry, particle's explanation
15:20 pmichaud (misread who was writing... hi particle!)
15:21 particle hi there pmichaud, glad you had safe travels
15:21 pmichaud but essentially, if we pre-compile a module to be a .pbc file, its subs have to have names that are unique from any other code we might be compiling/running
15:22 pmichaud otherwise the loaded pbc will overwrite any generated subs that happen to end up with the same (non-unique) names, or vice-versa
15:22 jhorwitz is this only for certain subs?  my concern is for subs like "handler" that mod_perl would need to call by name.
15:23 pmichaud it's primarily for anonymous blocks
15:23 pmichaud every block in p6 ends up becoming a parrot sub
15:24 pmichaud I suppose it's possible that we could try adding the :anon flag to blocks, and then changing the lookup code to use .const
15:24 particle pmichaud: why not use :anon...
15:24 particle heh
15:24 pmichaud particle:  the hard part is fetching the block if it doesn't have a name
15:24 pmichaud e.g.,   my $foo = { ... }
15:24 particle yep, gotta store a reference
15:24 pmichaud but getting to it via .const might work....
15:24 pmichaud I might try that.  That would be _much_ cleaner
15:25 particle i don't know that it would solve the problems, but it would come close
15:25 pmichaud if it does work, jhorwitz++ for asking the question :-)
15:25 particle that is, for non-anon generated subs
15:25 pmichaud non-anon generated subs already work okay so that's not really an issue
15:25 * jhorwitz lights the fires
15:26 particle but that could end up being user error....
15:26 pmichaud anyway, named subs always get names, yes.
15:26 particle do any anon generated blocks have loadinit?
15:26 pmichaud if they do, they'll still work
15:26 particle istr trouble with :anon and :load :init
15:26 particle perhaps that's been solved
15:26 pmichaud oh, you mean :load :init, as opposed to the new 'loadinit' attribute
15:27 particle actually, i think .sub 'onload' :anon :load :init works find
15:27 pmichaud we already do  :anon :load :init in lots of places
15:27 particle *fine
15:27 pmichaud I really like the :anon approach, I'll have to implement that.  Makes me very glad I didn't do it on the plane yesterday (because I'd be ripping out code today)
15:27 particle i wonder if grabbing handles to every block will help with exceptions, anyway
15:28 particle that's jit (jhorwitz-in-time) hacking for you
15:28 pmichaud that's also an excellent point (handles to every block)
15:29 pmichaud makes me wish I had a way of finding blocks other than via get_outer
15:29 particle yeah. hrmm.
15:30 particle find_sub only works for named subs
15:30 particle there's no api for finding every block
15:31 pmichaud oh, I was wishing for    .const 'Sub' outer = ....
15:31 pmichaud because it's then a compile-time op instead of a runtime one
15:32 Zaba joined #parrot
15:34 pmichaud I need lunch, though, then I'll do this.  and I probably need to feed kids.
15:37 pmichaud yes, kids are "starving".  bbiaw
15:44 davidfetter joined #parrot
16:03 Zaba joined #parrot
16:20 Zaba_ joined #parrot
17:01 Theory joined #parrot
17:33 dalek r30305 | coke++ | trunk:
17:33 dalek : [APL] minor update to get monadic.t passing again.
17:33 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30305
17:37 * smash wavez.
17:44 dalek r30306 | coke++ | trunk:
17:44 dalek : [APL] skip some failing tests. These used to work and they should be fixed.
17:44 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30306
17:50 desertmax_ joined #parrot
18:10 dalek r30307 | coke++ | trunk:
18:10 dalek : [languages] [t] temporarily remove some languages from the unified test harness.
18:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30307
18:12 cotto-work barney, ping
18:12 Ivatar joined #parrot
18:16 barney hi cotto
18:19 cotto-work how are you and francois planning on implementing print_r, serialize, etc?
18:19 dalek bernhard.schmalhofer@gmx.de | Pipp:
18:19 dalek link: http://www.perlfoundation.​org/parrot/index.cgi?pipp
18:20 cotto-work I've got working code in my phparray implementation for the functions (serizalize, var_dump, var_export and print_r), but I'm not sure if that's where they should go.
18:22 barney The builtin functions could delegate to the PMC methods. That looks cleaner than switches in the function itself.
18:22 cotto-work ok
18:22 cotto-work should that be the case for all types or just arrays and objects?
18:22 barney But I haven't really looked into functions yet. There is quite some basic syntax unsupported.
18:23 cotto-work it's still a wip
18:23 barney I'd say it should be consistent.
18:24 cotto-work so each type has a serialize, etc METHOD
18:24 Zaba joined #parrot
18:24 barney possibly using roles or inheritance
18:25 cotto-work that sounds good.  I don't think that any of those functions will be performance-critical (except perhaps serialize)
18:26 barney Performance doesn't matter at this stage. I think that all the optimisation we will ever need, can be stolen from Rakudo :=)
18:26 cotto-work premature optimization
18:26 purl premature optimization is the root of all evil or when you make the mistake of optimising for speed when really you should optimise for maintainability and correctness, in the first instance. or unnecessary optimization is premature by definition. or hacking with tequila in preference to wine
18:28 cotto-work agreed
18:31 jkva_ joined #parrot
18:34 dalek bernhard.schmalhofer@gmx.de | Pipp:
18:34 dalek link: http://www.perlfoundation.​org/parrot/index.cgi?pipp
18:36 tewk Is there a flag I can set to see the verbose make output, ie the full gcc cmdline
18:38 ruoso joined #parrot
18:41 barney tewk: don't think so. You could remove the @ after the tabs
18:42 particle iirc you can pass -n to see the commands that will run, but not actually run them
18:45 tewk tools/dev/cc_flags.pl looks to be the one hidding the info
18:45 tewk make -n helped
18:51 jkva_ Hmmm. Parrot Configure.pl dies when trying to determine the compiler to use. I have gcc installed (Ubuntu 7.10) - any tips? I'm sure I'm overlooking something.
18:54 dalek bernhard.schmalhofer@gmx.de | Pipp:
18:54 dalek link: http://www.perlfoundation.​org/parrot/index.cgi?pipp
18:54 cotto joined #parrot
18:55 barney jkva_: did you try a 'make realclean ; svn up ; perl Configure.pl '   ?
18:55 jkva_ I'll try. It's a fresh checkout, though.
18:56 jkva_ Nope, same thing. Grrr.
18:56 jkva_ "Compilation failed with 'cc'"
18:57 particle ok, so it's not finding gcc for you
18:57 particle perl Configure.pl --cc=/path/to/gcc
18:59 jkva_ Thanks, that seemed to work (and would make sense) but instead it now reports "Compilation failed with '/usr/bin/gcc'", drat. Perhaps I'm missing libraries
19:00 particle sudo apt-get install build-essentials # if i recall correctly
19:00 particle i've previously compiled parrot with gcc on ubuntu 7.10 32-bit, but i'm now running 8.04
19:01 jkva_ Ah, I see. --verbose=2 reports useful errors.
19:01 jkva_ And thanks for the build-essentials tip, I was trying to recall that name :)
19:02 particle yeah, it often slips my mind, too
19:04 jkva_ w00t, it works! Thanks so much :)
19:04 particle fabulous!
19:05 Zaba joined #parrot
19:06 jan joined #parrot
19:15 dalek r30308 | julianalbo++ | trunk:
19:15 dalek : added 'f' alias for 'script' command in debugger, that was listed in help but not implemented
19:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30308
19:21 jkva_ left #parrot
19:29 Zaba_ joined #parrot
19:35 Zaba joined #parrot
19:35 dalek r30309 | julianalbo++ | trunk:
19:35 dalek : added '(s) stack' command to debugger, it was listed in help but not implemented
19:35 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30309
19:38 Tene Today has been insane and exhausting, and it's only half-over.  I don't think I'm getting anything useful done tonight after work.
19:39 Andy joined #parrot
19:44 NotFound Someone wants to play a little with parrot_debugger?
19:45 Casan joined #parrot
19:52 rurban joined #parrot
19:52 dalek rurban@cpan.org | Parrot Development on Windows:
19:52 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?parrot_development_on_windows
19:53 nopaste "rurban" at 212.183.81.9 pasted "check config_hash key?" (18 lines) at http://nopaste.snit.ch/13832
19:54 rurban Anybody knows why I get a VM PANIC when I access IGLOBALS_CONFIG_HASH?
20:13 integral joined #parrot
20:14 NotFound rurban: Why do you call VTABLE_elements on a PMCNULL?
20:14 rurban config_hash is populated from VTABLE_get_pmc_keyed_int
20:15 NotFound rurban: after
20:15 rurban Just to be sure I initialized it with PMCNULL.
20:16 nopaste "rurban" at 212.183.81.9 pasted "check config_hash key?" (16 lines) at http://nopaste.snit.ch/13833
20:18 nopaste "rurban" at 212.183.81.9 pasted "full 56996-fhs-runtime.patch" (216 lines) at http://nopaste.snit.ch/13834
20:20 NotFound Did you have a backtrace?
20:22 rurban No. Just a coredump which is not really usable on cygwin. Should I try on linux?
20:22 NotFound I can't try your patch at this moment, sorry.
20:23 donaldh joined #parrot
20:38 dalek r30310 | julianalbo++ | trunk:
20:38 dalek : new --script option in parrot_debugger, same as the 'script' debugger command line
20:38 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30310
20:43 Limbic_Region joined #parrot
20:43 smash joined #parrot
20:53 Whiteknight joined #parrot
21:01 dalek r30311 | Whiteknight++ | gsoc_pdd09:
21:01 dalek : [gsoc_pdd09] update to trunk from r30276
21:01 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30311
21:04 AndyA joined #parrot
21:04 dalek r30312 | Whiteknight++ | gsoc_pdd09:
21:04 dalek : [gsoc_pdd09] add patch from chromatic++ that helps prevent garbage from being needlessly inserted into newly-created PMCs. Also, remove a defunct comment.
21:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30312
21:07 rurban who does tcl?
21:07 rurban I got minor test failures for float inexacts. #167,#202,-#204 deviation in the last digit
21:11 Whiteknight Particle was doing TCL, I think
21:11 Whiteknight or, he was one of them
21:11 particle nope, that's dietcoke
21:12 Whiteknight oops, my mistake
21:12 rurban That's tcl/t/cmd_expr.t #167,#202,#203,#204
21:16 donaldh ping jhorwitz
21:16 purl I can't find jhorwitz in the DNS.
21:16 rurban hmm, lib/test_more.tcl needs some tests for such float weirdness
21:16 donaldh jhorwitz: ping
21:16 jhorwitz donaldh: ping
21:17 donaldh jhorwitz: I'm trying to use mod_parrot with perl6 and getting an error during apache startup.
21:17 jhorwitz join the club
21:17 donaldh jhorwitz: Use of contextualizer :: not implemented. at line 35, near "::handler("
21:17 jhorwitz LOL
21:17 jhorwitz you need to apply the included patch
21:17 donaldh expected then
21:18 donaldh ah, okay.
21:18 jhorwitz however...
21:18 purl Consequently, and fuuuuurther-moooooore
21:18 * jhorwitz hates purl even more now
21:18 jhorwitz the patch will fail on io.pir, cuz those changes are already in trunk
21:18 jhorwitz so just revert that file after you apply the patch
21:19 donaldh That'll be why I ignored the patch earlier. I noticed that io.pir had changed, but didn't check other files.
21:19 nopaste "rurban" at 212.183.81.9 pasted "tcl float inexactness in cygwin" (25 lines) at http://nopaste.snit.ch/13835
21:19 jhorwitz donaldh: yeah, you need interpolated namespaces, which isn't in rakudo just yet
21:19 donaldh gotcha.
21:20 donaldh jhorwitz: thanks, I'll try that now.
21:21 rurban 0.7.0 makes me happy. almost all tests pass. 99.98% for parrot (2 failed), 100% for perl6, 98.98% (25 failed) for the other languages
21:23 donaldh 0.7.0 ? Is that trunk latest?
21:24 rurban well, trunk of course
21:24 rurban r30309
21:25 donaldh k
21:25 dalek r30313 | smash++ | trunk:
21:25 dalek : [ports/debian]: updates from last build
21:25 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30313
21:30 smash debian packages for 0.6.4 now available, just mailed p2p
21:30 particle yay, just in time for 0.7.0 tomorrow :)
21:30 smash i'll build new ones as soon as 0.7.0 is available
21:31 smash a couple of things still need to be fixed, so maybe not tomorrow.. but sometime this week
21:31 particle that'll be lovely. smash++
21:32 rurban smash: How did you solve the /usr/runtime problem? #56996
21:33 rurban I mean, did you test the binary package with renaming your build-dir?
21:40 Zaba joined #parrot
21:42 donaldh Wow, my linkstation is slooooow. Nearly 8 minutes to generate gen_actions.pir
21:46 Whiteknight wow, that is slow
21:47 donaldh It only has a 400MHz ARM I suppose.
21:48 cotto_work you can't expect too much from a machine like that
21:48 cotto_work otoh, you won't run out of space. ;)
21:48 donaldh troo
21:49 smash rurban: i wasn't even aware of that issue, but i'm seeing it's still looking in runtime*
22:15 dalek r30314 | Whiteknight++ | gsoc_pdd09:
22:15 dalek : [gsoc_pdd09] remove some obnoxious debugging messages that weren't helping, comment out the PMC sweep code, which exposes a different problem.
22:15 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30314
22:55 StephenPollei joined #parrot
23:00 Tene Any parroters in Austin, TX?
23:03 Limbic_Region Tene - http://austin.pm.org/ may be worth checking out
23:03 Tene Oh, clever.
23:04 Limbic_Region pmichaud is over 200 miles away - he is the only parroter I know for sure in TX
23:18 tetragon joined #parrot
23:23 davidfetter joined #parrot
23:25 pmichaud PerlJam has been Austin-based
23:26 pmichaud austin.pm is mildly active, yes.
23:26 pmichaud I did a parrot presentation that was co-sponsored by austin.pm
23:27 pmichaud it's probably about time for me to do another :-)
23:37 dalek r30315 | Whiteknight++ | trunk:
23:37 dalek : [PDD] a few clarifications to PDD09, some better wording, spelling fixes. Expand on some topics that didn't have great coverage before.
23:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=30315

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

Parrot | source cross referenced