|
01:10
JimmyZ joined
|
|||
| JimmyZ | jnthn: Bump dyncall version # I think we didn't update to a new dyncall version :P | 01:10 | |
|
01:14
FROGGS_ joined
06:03
Ven joined
06:59
cognome joined
07:00
FROGGS joined,
zakharyas joined
07:06
leedo joined,
cognome_ joined
|
|||
| sergot | hi oi/ | 07:07 | |
| o/ | |||
| nwc10 | \o | ||
|
07:45
cognome joined
|
|||
| FROGGS | alt grß+o | 07:47 | |
|
08:00
PerlJam joined
09:25
leont joined
|
|||
| dalek | arVM: 8153f38 | jonathan++ | docs/ChangeLog: Seems we didn't bump dyncall version after all. |
10:34 | |
|
10:47
woolfy1 joined
11:05
brrt joined
|
|||
| brrt | \o | 11:07 | |
| nwc10 | o/ | 11:08 | |
| brrt | i'm in a lecture now :-) | 11:09 | |
| timotimo | yays uni | 11:11 | |
| brrt | very | ||
|
11:40
FROGGS joined
|
|||
| jnthn | so study... | 12:07 | |
| OK, so apparently MoarVM is incredibly bourgeois | |||
| I am missing a load of commits on my Linux VM | 12:08 | ||
| Which is where I cut release. | |||
| And...4 attempts to git pull later, it never manages to grab all the objects :/ | |||
| timotimo | what the ... | ||
| jnthn | 3.00 KiB/s :/ | 12:09 | |
| wtf. | |||
| This hsa gotta be some kind of weird traffic shaping. :/ | |||
| timotimo | must be | 12:10 | |
| jnthn | hmm, wonder what protocl it's using... | ||
| ah, https | |||
| of course anybody using https must be doing something terrible... :) | |||
| timotimo | aye | 12:11 | |
| jnthn | Yup. Use git protocol and it's done already. | ||
| Wow. | |||
| The craziest thing is that if I tunnel my HTTP and HTTPS traffic via Germany over SSH, it still often works out faster than direct requests from here... | 12:13 | ||
| ...and I get German ads all the time :P | |||
| FROGGS | Weißbier \o/ | ||
| jnthn | darn | ||
| the submodules are https | |||
| FROGGS | I think there is a ticket for that already :o) | 12:16 | |
| jnthn | Of course submodules suck so I'm having fun trying to imrpvoe things | ||
| FROGGS | :/ | 12:17 | |
| jnthn | Of course, the alternatives to using them typically suck just as hard... | ||
| Does anybody want to cut a MoarVM 2014.09 tarball? :) | |||
| FROGGS | jnthn: gimme instructions :o) | ||
| jnthn | FROGGS: docs/release_guide.md | 12:18 | |
| FROGGS: I did 1..3 | |||
| FROGGS | taht's too easy | ||
| jnthn | So basically make release VERSION=2014.09 | ||
| And you get a tarball | |||
| Make sure you git pull'd and configure'd first and that's all it needs. | |||
| If you fancy doing step 5 and 6 too, feel free, but I'm happy to also. :) | 12:19 | ||
| I think I might do dinner first though...kinda hungry. :) | |||
| FROGGS | k, I'll report here | 12:20 | |
| jnthn | Cool. Just drop the tarball somewhere and lemme know what step you go to, and I'll do the reset once I'm back. | ||
| FROGGS++ | |||
| *rest | 12:21 | ||
| bbl & | |||
| FROGGS | "This is MoarVM version 2014.09" | 12:22 | |
| 4, 5 and 6 done... froggs.de/perl6/MoarVM-2014.09.tar.gz | 12:26 | ||
| lizmat | FROGGS++ jnthn++ | 12:27 | |
| masak | FROGGS++ | 12:41 | |
| jnthn | Now at www.moarvm.org/releases/MoarVM-2014.09.tar.gz | 13:41 | |
| FROGGS++ # tanks! | |||
| *thanks! | |||
|
13:42
Ven joined
|
|||
| FROGGS | snacks! | 13:43 | |
| :D | |||
|
13:45
Ven joined
13:48
vendethiel joined
14:04
JimmyZ joined
|
|||
| dalek | href="https://moarvm.org:">moarvm.org: 5963e3d | jonathan++ | / (4 files): Update website for 2014.09. |
14:12 | |
| nwc10 | jnthn: are you building that tar.gz on linux? | 14:13 | |
| jnthn | nwc10: yes | ||
| Usually :) | |||
| Today's attempt ended up with delegation :) | 14:14 | ||
| nwc10 | jnthn: you're aware that it's possible to beat gzip by about 5%? | 14:15 | |
| perl5.git.perl.org/perl.git/blob/v5...e.pod#l836 | |||
| jnthn | nwc10: No. I guess being aware of such things is what gets you a 1400 line release guide rather than a 40 line one... ;-) | 14:19 | |
| nwc10 | that's not *totally* fair. I think the bigger problem is having a festering mess that barely hangs together, and a lot of silent-or-grumpy downstream that expect perfection (or better) | 14:20 | |
| and only report bugs about 1-5 years after you released them | 14:21 | ||
| jnthn | Well, it's totally unfair 'cus MoarVM is just a VM, so you'd need to add the NQP and Rakudo guides too :) | ||
| How much applies for those doing the monthlies, ooc? | 14:22 | ||
| nwc10 | I don't know. I think "quite a bit" just to keep everything in good shape | 14:23 | |
| nine | anyway just releasing a tar.xz instead of a tar.gz seems simple enough | 14:26 | |
| nwc10 | In Perl 5 land we didn't think that "intead" was great | ||
| currently we're doing bz2 and gz | 14:27 | ||
| I think it might be a plan to go to xz and gz | |||
| jnthn | I'm not sure MoarVM is currently getting downloaded enough to worry a lot about this... ;) | ||
| nwc10 | also, you don't have 200+ mirrors | ||
| jnthn | Right :) | ||
| Another curious little data point: | 14:28 | ||
| rakudo-star-2014.08-parrot.msi 03-Sep-2014 14:25 17M | |||
| rakudo-star-2014.08-moar.msi 03-Sep-2014 15:00 9.5M | 14:29 | ||
| tadzik | heh | ||
| jnthn | That's quite a size difference... | ||
| tadzik | and has moar features | ||
| jnthn | Yes :) | ||
| tadzik | ba-dum-tshh | ||
| jnthn | :P | ||
|
14:33
JimmyZ_ joined
|
|||
| carlin | so, moar is less | 14:34 | |
| jnthn | moar or less... | 14:35 | |
| carlin | less is moar | ||
|
14:37
FROGGS[mobile] joined,
Ven joined
16:10
Ven joined
16:36
FROGGS joined
16:53
Ven joined
17:02
zakharyas joined
17:41
brrt joined
17:47
Ven joined
18:01
camelia joined
18:10
dalek joined
|
|||
| brrt | timotimo: ping? | 19:17 | |
|
19:18
Ven joined
|
|||
| dalek | arVM/cglobal: c1c22e9 | (Tobias Leich)++ | / (2 files): conditionally use/install our libtommath headers |
19:18 | |
| MoarVM/cglobal: a924435 | jonathan++ | src/6model/serialization.c: | |||
| MoarVM/cglobal: Don't try to serialize file handles. | |||
|
19:19
dalek joined
|
|||
| brrt | this dude is going broke real soon: medium.com/life-at-primeloop/putti...757946d9fe | 19:19 | |
|
19:20
daxim__ joined
19:21
zakharyas joined
|
|||
| brrt | 'falling back on mutable state and for loops make me feel like a bad programmer' - or a wuss :-p | 19:35 | |
| timotimo | pong | 19:44 | |
| brrt ^ | 19:46 | ||
|
20:08
brrt joined
|
|||
| brrt | \o timotimo | 20:17 | |
| who killed camelia again | |||
|
20:17
camelia joined
|
|||
| lizmat | diakopter++ working on issues on host08 | 20:20 | |
| diakopter | [Coke] is I think | ||
|
20:21
camelia joined
|
|||
| [Coke] | I'm stabbing at things blindly. | 20:22 | |
| I do think we want to decouple the rakudo build process into 3 separate builds. having rakudo-moar wait for rakudo-jvm to build is bad. | |||
| dalek | arVM: 699360b | (Tobias Leich)++ | / (8 files): add op nativecallglobal |
||
|
20:27
brrt left
|
|||
| timotimo | brother: what was up? :) | 20:34 | |
| er | |||
| brrt is already gone again :\ | |||
| ah, what the next step for the jit should be | |||
| don't really know ... | |||
| at some point we're going to want to throw away boxing/unboxing - and of course the dynasm change that'd let us do "proper" register allocation | 20:35 | ||
| and then at some point we'll want common subexpression elimination and loop invariant hoisting ... | 20:39 | ||
| and if we can have a "no smallbigint path" version of the bigint ops, we could inline the smallbigint computations for all the ops into the generated assembly and call to the only-bigint-path version of the op when we discover it's not "compatible" | 20:40 | ||
|
21:27
jlaire joined
21:29
avuserow joined
22:03
leont joined
|
|||
| [Coke] | j: say 3; #not building anything right now | 22:44 | |
| camelia | rakudo-jvm 36bada: OUTPUT«Can't call method "syswrite" on an undefined value at /home/p6eval/jvm-rakudo/eval-client.pl line 32.» | ||