Camelia, the Perl 6 bug

IRC log for #moarvm, 2013-08-02

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

All times shown according to UTC.

Time Nick Message
00:24 benabik joined #moarvm
00:41 ggoebel2 joined #moarvm
00:43 ggoebel joined #moarvm
00:44 ggoebel3 joined #moarvm
01:15 ggoebel joined #moarvm
01:38 JimmyZ joined #moarvm
01:44 ggoebel2 joined #moarvm
01:45 ggoebel3 joined #moarvm
01:47 ggoebel4 joined #moarvm
01:49 ggoebel5 joined #moarvm
01:51 JimmyZ joined #moarvm
02:18 cognominal joined #moarvm
02:33 JimmyZ_ joined #moarvm
02:34 JimmyZ_ joined #moarvm
03:28 japhb joined #moarvm
05:11 birdwindupbird joined #moarvm
07:35 dalek MoarVM/readlineintfh2: 5efa873 | jimmy++ | 3rdparty/linenoise/linenoise.c:
07:35 dalek MoarVM/readlineintfh2: added ctrl-l, clear screen support for Windows
07:35 dalek MoarVM/readlineintfh2: review: https://github.com/MoarVM/MoarVM/commit/5efa87397f
07:45 FROGGS joined #moarvm
07:47 FROGGS o/
08:29 dalek MoarVM/readlineintfh2: d4aca36 | jimmy++ | 3rdparty/linenoise/linenoise.c:
08:29 dalek MoarVM/readlineintfh2: added ctrl-w, delete previous word support for Windows
08:29 dalek MoarVM/readlineintfh2: review: https://github.com/MoarVM/MoarVM/commit/d4aca365af
09:00 diakopter JimmyZ: on windows will you enable the traditional Ctrl-z semantics (end of stdin)?
09:19 JimmyZ diakopter: on linux doesnot support ctrl-z yet too
09:20 diakopter I don't understand that sentence
09:21 JimmyZ diakopter: linenoise doesn't support ctrl-z
09:23 JimmyZ diakopter: so I won't make it works on windows too..
09:23 JimmyZ unless we really need it, I think
09:24 diakopter ok; I was merely curious if you had planned on it
09:26 JimmyZ fetures that moarvm is used on linux, and I make it works on windows also, it's my plan. and I think it's done now :P
09:27 diakopter I'm afraid I can't understand that sentence either. :(
09:27 JimmyZ so for now, just waiting for jnthn ...
09:27 FROGGS diakopter: he just makes features available that are support on linux by default
09:27 JimmyZ diakopter: sorry, that must be becase  of my poor english :(
09:28 diakopter JimmyZ: also, I'm quite obtuse at 2:30 a.m.
09:28 FROGGS diakopter: ...that are supported by linenoise on linux be default
09:28 diakopter (unfortunately)
09:28 FROGGS by*
09:28 FROGGS me
09:28 FROGGS can't type either
09:28 JimmyZ diakopter: you are working hard :P
09:28 diakopter JimmyZ: or hardly working
09:29 JimmyZ hehe
09:31 JimmyZ FROGGS: not really, I added ctrl-r for linux too :P
09:31 FROGGS ha! gotcha!
09:32 diakopter :)
09:40 JimmyZ [root@ nqp-cc]# echo 'hello world' | nqp nqp-moar-cc.nqp test.nqp
09:40 JimmyZ 11
09:40 JimmyZ hello world
09:40 JimmyZ [root@nqp-cc]#
09:40 JimmyZ no segfault
09:40 diakopter cat test.nqp
09:41 diakopter uh oh, running unstable software as root... you'll format your disk accidentally :)
09:43 diakopter JimmyZ: what's in test.nqp?
09:43 JimmyZ diakopter: https://gist.github.com/zhuomingliang/6119449
09:44 diakopter I guess readlineintfh removes a line break?
09:44 * JimmyZ doesn't know
09:45 diakopter well if say() doesn't add another one...
09:47 JimmyZ yeah
09:47 JimmyZ it removes one
09:47 JimmyZ actually, it doesn't insert 'enter'
09:47 diakopter oh
09:48 diakopter but \r\n on stdin...
09:48 diakopter (end a line with ^ on windows)
09:48 JimmyZ it's easy to insert "\n", if we want
09:49 * JimmyZ didn't see a '^' on windows
09:50 diakopter ?
09:51 diakopter what do you mean "didn't see"
09:51 diakopter *sigh* troll on my grant app page is a troll
09:53 diakopter (don't anyone reply please.. he contraverted my affirmation of the value of his opinion)
09:53 arnsholt That last one is a real WTF
09:53 JimmyZ diakopter: re: but \r\n on stdin... (end a line with ^ on windows), I don't see there is a '^'
09:53 diakopter what do you mean you don't see?
09:53 arnsholt A badly expressed echo of chromatic's usual tirade
09:53 diakopter where don't you see it?
09:54 JimmyZ who is Carlos del Rey ?
09:54 diakopter personally, I don't have a problem with chromatic's critiques, except their style occasionally
09:54 JimmyZ diakopter: on cmd
09:55 diakopter JimmyZ: where on cmd?
09:55 diakopter just type a ^ then enter
09:57 diakopter C:\> echo foo foo ^ENTER bar bar |more
09:57 diakopter foo foo
09:57 diakopter bar bar
09:58 diakopter arnsholt: yeah, my imagined reply would be something like this:
09:58 JimmyZ hmm, I think I have some problems with english ...
09:58 diakopter 1. When was this called Blizkost? Blizkost never wanted to handle anything more than Parrot.
09:59 diakopter 2. You're a vanity VM
09:59 diakopter 3. Parrot already works?
09:59 diakopter 4. Perl-6 is a long farce? Just wait to see how long it can get!
10:00 JimmyZ Before a successful genius are nuts
10:00 JimmyZ ^^ is what I want to reply :P
10:01 diakopter 5. It's not always fun to start something new. It's often quite painful and annoying, but worth it in the end. It's tough to start over because it's admitting that the work of people you care about [often yourself if it's your own work] would be best improved by starting over, because of huge fundamental problems
10:01 arnsholt While it would be kind of amusing if you made that reply, I think not troll-baiting is a good idea, yeah =)
10:05 diakopter 6. re "my money TPF is going to throw away" - you know what, you sorry excuse for a troll?  It's not like I wouldn't do this work anyway if the grant was approved, so the nod from TPF would simply be "recognition" for the work, and encouragement to do more such work if it was successful. You think I'd choose to accept another gig for those hours of my time at that rate, at a 70% pay cut?  The time is already worth far more to me than the rate ...
10:05 diakopter ... of compensation.
10:05 diakopter er s/was/wasn't/
10:05 diakopter [he's not a very good troll]
10:11 diakopter FROGGS: yes, I was joking that I would actually reply with these points. I'm being facetious/mocking and tongue-in-cheek, partly.
10:12 diakopter [I don't actually think Carlos is a vanity VM]
10:15 diakopter I would be glad to discuss all of Carlos' points civilly on channel or telephone or in person or whatever. :) I'm quite confident that would go quite well [for everyone].
10:15 nwc10 "Parrot works? - well, MoarVM is pretty much the implementation of the Parrot developers' own Lorito plan, only without the PIR layer on top"
10:18 diakopter I tried to discover what the Lorito plan was... but didn't get further than "reduced instruction set VM"
10:18 diakopter .. which MoarVM certainly isn't
10:18 nwc10 ah OK. then I'm confused
10:18 nwc10 acutally, it sort of feels like it is
10:19 nwc10 its "reduced" compared with parrot's current instruction set
10:19 tadzik hah, this guy
10:19 tadzik to me, he lost it all at "LOL"
10:20 diakopter nwc10: have you seen oplist lately? :) it's only 1/4 the count of parrot's ops because it doesn't have specializations of every op for various types of args. Instead of built into each op, the coercion of args to the ops is built into the QAST->MAST compiler, so it doesn't need the specialized editions
10:20 tadzik diakopter | [I don't actually think Carlos is a vanity VM]
10:20 tadzik 'course not. His mom is.
10:21 diakopter er, s/args/operands/g
10:23 diakopter nwc10: I think of NQP's huge oplist as the analogue of the entire set of JVM's rt.jar callouts to native functions.. just in one flat namespace instead of how JVM names them
10:23 diakopter you know, several hundred native function interfaces... doing nativey things
10:24 diakopter HOWEVER :D :D :D
10:26 diakopter if *I* were making a reduced instruction set VM for dynamic languages, I would make one that didn't include function calls at all, but instead implemented function calls with lower primitives, where bytecode frames were objects that a goto op could goto into.
10:27 diakopter .. which is actually *almost* how moarvm does it, except the frame objects are not accessible in the HLL b/c they're not GC'd
10:29 diakopter .oO( someday some insane [read: one of mine] branch of moarvm repo will have all VM-internal objects accessible to HLL... )
10:48 benabik_ joined #moarvm
11:36 colomon joined #moarvm
12:21 JimmyZ Good evening
12:22 FROGGS hi JimmyZ
12:32 JimmyZ Hi FROGGS :P
12:52 masak greetings JimmyZ
12:55 JimmyZ \o/ 麦高
12:59 masak \o/
13:35 diakopter nwc10++ # excellent questions
13:37 diakopter nwc10: "Until distributions pick up MoarVM to package, people wanting to use it will have to build it from source"  well, as mst said on #perl6 the other day, the idea is to have a Rakudo::MoarVM package that builds/installs the whole thing using cpan/m
13:40 diakopter nwc10: re 1., I wouldn't request a halfway mark payment
13:42 diakopter nwc10: re 2., I think it'd be good for acceptance tests to use 5.18 and .20 only.. but where I'd also be aiming for 5.8. I strongly suspect that any bugs fixed between 5.8->now would not prevent the interop from working at all, but instead might affect some small aspect of it.
13:44 diakopter the only thing that's really p5 C-heavy is the stash sync/walk thingie
13:44 diakopter the rest is baby p5/p5C
13:46 diakopter nwc10: re 3., I think what you described in this paragraph would be a great test that's pretty decisive: "Something like having a web framework running in Perl 6 accessing databases via DBIx::Class and DBI running on Perl 5? Calls from Perl 6 into methods written in Perl 5, passing callbacks to Perl 6 routines (and vice versa)."
13:46 diakopter nwc10: however, the part about inter-VM class derivation seems less certain
14:00 FROGGS joined #moarvm
14:05 masak nwc10++ # comment
14:06 masak I was very tempted to state for any blog comment readers that Carlos del Rey is a troll, just to provide some context. but I think I'll skip that too until further notice.
14:09 masak I just hope that all the "+1" comments contrast enough with his repetitive colloquialisms to show that he's not exactly worth paying attention to.
14:18 diakopter No. 3 becomes more certain when I admit the possibility of integrating with Moose/Class::MOP (and Moo and the rest)
14:18 diakopter (with new releases of those and/or plugins)
14:19 diakopter wouldn't want to try to make it work with default p5 inheritance
14:19 diakopter afk all day &
15:29 benabik joined #moarvm
15:48 Alpha64 joined #moarvm
16:10 JimmyZ_ joined #moarvm
16:10 benabik joined #moarvm
16:46 Alpha64_ joined #moarvm
16:48 Alpha64__ joined #moarvm
17:41 colomon joined #moarvm
17:47 colomon joined #moarvm
17:49 diakopter all day was short
17:54 benabik joined #moarvm
17:59 lizmat 3.5 hours, wow!
18:01 diakopter lizmat: do you agree about emphasizing/demanding moose for MOP-level interop?
18:01 lizmat Moose seems sensible to me, as that is backported to quite a few perls ago
18:02 lizmat MOP is much more recent: demanding that would perhaps cut you off from older perls ?
18:02 diakopter I just meant Class::MOP
18:02 diakopter not p5 MOP
18:06 lizmat ah, I have no idea how far back Class::MOP is supported
18:06 lizmat if it is far enough for you, then demanding it seems like a small price to pay
18:53 ggoebel5 joined #moarvm
19:00 ggoebel joined #moarvm
19:25 _ilbot joined #moarvm
19:25 Topic for #moarvm is now https://github.com/moarvm/moarvm | IRC logs at  http://irclog.perlgeek.de/moarvm/today
19:28 ggoebel2 joined #moarvm
19:40 hoelzro joined #moarvm
19:57 ggoebel joined #moarvm
20:20 crab2313 joined #moarvm
20:26 ggoebel joined #moarvm
20:42 ggoebel joined #moarvm
20:47 bronco_creek joined #moarvm
21:24 cognominal joined #moarvm
22:12 FROGGS joined #moarvm
23:33 benabik joined #moarvm
23:51 benabik joined #moarvm
23:54 crab2313 joined #moarvm

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