Camelia, the Perl 6 bug

IRC log for #moarvm, 2013-07-01

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

All times shown according to UTC.

Time Nick Message
00:55 FROGGS joined #moarvm
01:12 FROGGS_ joined #moarvm
02:20 ggoebel joined #moarvm
03:17 woosley is MoraVM broken for now?
03:17 woosley make test
03:17 woosley prove -e /home/woosley/code/MoarVM/nqp-cc/install/bin/nqp t/moar t/qast
03:17 woosley t/moar/6model.t ................. Bytecode validation error: instruction operand type does not match register type
03:17 woosley t/moar/6model.t ................. 1/2 Bytecode validation error: non-existent operation bank 5 op 46
03:17 woosley t/moar/6model.t ................. Failed 2/2 subtests
03:17 woosley t/moar/annotated.t .............. Bytecode validation error: non-existent operation bank 5 op 46
03:17 woosley t/moar/annotated.t .............. Failed 1/1 subtests
03:17 woosley t/moar/associative.t ............ ^Cmake: *** [test] Interrupt
03:17 sorear o_O
03:23 woosley OK, maybe I did not make again after I did 'git pull' :)
03:26 sorear is it better now?
03:30 woosley /usr/bin/ld: cannot find -luuid
03:30 woosley collect2: ld returned 1 exit status
03:30 woosley make: *** [moarvm] Error 1
03:31 woosley hmmm, I have it installed, but still see  this error,
03:34 sorear Do you have the dev library?  (if appropriate for your OS)
03:39 FROGGS joined #moarvm
03:52 woosley hooray, I got it fixed and looks like tests are passing now
03:55 sorear \o/
04:41 FROGGS joined #moarvm
04:48 birdwindupbird joined #moarvm
06:31 FROGGS joined #moarvm
09:17 * jnthn had 'em passing yesterday also :)
09:57 cognominal joined #moarvm
13:18 birdwindupbird joined #moarvm
13:26 Guest1337 joined #moarvm
15:12 benabik joined #moarvm
15:14 benabik ~~
15:15 BinGOs Smartmatch is an experimental feature at benabik
15:16 benabik Indeed.  Smartmatching benabik is highly experimental.  Tends to be buggy without coffee.
15:16 JimmyZ remove sink context, per sorear++ said :P
15:16 JimmyZ opps
15:16 JimmyZ I meant next window
15:17 BinGOs #perl6 is 28 windows that <- way
15:17 benabik IRC is non-euclidian.  It's 1 tab that -> way for me.
15:31 [Coke] benabik: weird. it's in the same place on my irc.
15:39 FROGGS joined #moarvm
15:40 diakopter jnthn: did you see my note a week (or few?) ago about using something from the HLL config for initializing all pointers instead of NULL?
15:41 jnthn diakopter: No, missed that. Hmm.
15:42 diakopter (like the Any object or whatever)
15:42 jnthn Interesting idea.
15:42 jnthn diakopter++
15:42 jnthn Oh...
15:42 jnthn But it doesn't quite work.
15:42 jnthn We still gotta have nqp::isnull(...), nqp::ifnull(...) work
15:42 diakopter well it wouldn't work until enough is loaded
15:42 jnthn And there are places where it matters a lot to distinguish Any/Mu from null.
15:43 diakopter hm, oh.
15:43 jnthn So, interesting but I think we can't do that without breaking lots of things.
15:43 diakopter what things use that?
15:44 jnthn It's used in auto-viv for one, also array.delete. Probably as a kind of sentinel in various other situations.
15:44 diakopter my intuitive guess is that those could use another value
15:44 jnthn well, we get it using nqp::null()
15:44 diakopter yeah, that. :)
15:45 jnthn I think the other thing, though, is that we tend to want to be told when a null value leaked in.
15:45 jnthn It's usually an implementation error.
15:45 jnthn And maybe it's better that those blow up sooner.
15:47 diakopter jnthn: remind me: does a Perl 6 array have each slot initialized with an empty Scalar object?
15:47 jnthn diakopter: On first write
15:48 jnthn my @a; @a[4] = 10; # Scalar created due to the assignment
15:48 diakopter jnthn: do we have a smart "empty scalar" object to avoid duplication?
15:49 diakopter (when something else tries to bind or pull from it, it auto-vivifies to another empty one, I guess)
15:49 jnthn diakopter: I'm not quite sure I follow...
15:49 jnthn Oh, maybe I do...
15:50 diakopter oh, the array repr doesn't ever bulk create empty scalar objects?
15:50 jnthn No
15:50 diakopter oh, ok
15:50 diakopter nm
15:50 jnthn They come about "on demand"
15:50 jnthn Note that a List doesn't need to create them
15:50 jnthn $*IN.lines.map(...).grep(...) is List all the way through rather than Array.
15:52 diakopter oh
15:52 diakopter which repr does List use?
15:53 jnthn List itself is a P6opaque; it holds a pair of VMArray
15:53 jnthn For reified and unreified portions respectively
15:53 jnthn Or something like that :)
15:54 diakopter interesting
18:53 ggoebel2 joined #moarvm
19:02 lizmat joined #moarvm
19:21 tomyan joined #moarvm
20:19 tomyan joined #moarvm
22:14 FROGGS joined #moarvm
22:24 FROGGS joined #moarvm
22:58 FROGGS joined #moarvm

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