00:00
kjs_ joined
00:43
llfourn joined
01:44
llfourn joined
02:40
llfourn joined
02:47
ilbot3 joined
06:54
virtualsue joined
07:38
kjs_ joined
07:50
domidumont joined
07:56
domidumont joined
11:07
virtualsue joined
11:15
vendethiel joined
12:22
zakharyas joined
13:10
brrt joined
14:25
lucasb joined
|
|||
lucasb | hello o/ | 14:31 | |
If the 3rdparty dependencies dynasm, dyncall, libuv are so rarely updated, what's the benefit of having them as git submodules? Bundling the real source code like the other dependencies wouldn't simplify the build process a little? I mean, when building from a fresh git clone, there would be no need to download them over the net. | 14:32 | ||
timotimo | we have a few patches in those projects :( | ||
lucasb | Aren't they mantained in the repos MoarVM/dyncall and MoarVM/dynasm? I meant just to copy the source from those repos, not the offical upstream repos. | 14:34 | |
brrt | hi lucasb | ||
lucasb | o/ timotimo and brrt | ||
brrt | the dynasm 'fork' is maintained by us, yes, i don't know about dyncall, libuv | ||
libuv we want to keep updated, but it's complex | |||
anybody volunteering to get libuv 1.7 or something working on moarvm would be welcome | 14:35 | ||
lucasb | I just say this, because I was playing with my ad-hoc build script... and was getting tired of that "updating submoules" step :) | ||
JimmyZ | well, dyncall is in git repo, because it has no git repo | 14:36 | |
lucasb | yeah, they seem to use mercurial... hg.dyncall.org/pub/dyncall/dyncall | 14:37 | |
dalek | arVM: 07906c9 | timotimo++ | src/spesh/optimize.c: take care when optimizing istrue to unbox_i |
||
JimmyZ | anyway , using git submodules is not bad | 14:38 | |
brrt | timotimo++ that's fast | ||
lucasb | brrt: I think flussence was the one building moarvm with an updated libuv, and not having found any problems | ||
timotimo | it was very easy, since i knew what was up :) | ||
brrt | when was that? | ||
libuv churn is significant | 14:39 | ||
well, thats why we've hired you... oh | |||
timotimo | :D | ||
brrt | anyway, i'm all pro a libuv update | 14:50 | |
but i couldn't make it work last time | 14:51 | ||
and i don't know why | |||
timotimo | it's probably a good idea to get on latest libuv soon-ish; from a security perspective, too! | ||
brrt | hmmpf | ||
i'm willing to make a branch | 14:52 | ||
timotimo | why are you wasting your precious new-jit-skills on libuv janitoring? :) | 14:53 | |
brrt | because... i can't really focus on the new jit right now, what with the thesis work i'm doing | ||
timotimo | oh, so since you don't have enough concentration to bring the jit forward, you'll spend time on a "low effort" thing like bumping libuv version | 14:54 | |
brrt | kind of, yes | ||
:-) | |||
timotimo | i shouldn't tell you what to do with your time :) | 14:55 | |
brrt | no offense was taken :-) | ||
on the other hand, you bring up a good point | 14:56 | ||
shouldn't linger on too much on it | |||
i've decided on a number of things, though | |||
one, that i want to specify the register (type) and possibly the 'used register' set in tiles | 14:57 | ||
e.g. | |||
timotimo | sorry, i'm about to go AFK for ~30 minutes | ||
brrt | i want the div tile, if any, to specify that it needs rax,rcx, and yields rax, whereas mod tile also needs rax, rcx but yields rcx | ||
e.g. i need a 32 bit register flag set for describing which tiles are 'fixed' to certain registers | 14:58 | ||
brrt wonders whether it is worth it to install virtualbox, since it works so much better and easier than qemu | 15:18 | ||
JimmyZ | as long as you don't use desktop, it is worth | 15:23 | |
I mean in the guest os | 15:24 | ||
flussence | `ldd /usr/bin/moar` ==> Ā«libuv.so.1 => /usr/lib/libuv.so.1Ā» here :) | 15:29 | |
(though I'm probably not the person to ask if it's not working somewhere. I just flipped the switch on one day and it worked first time, so I kept it) | 15:34 | ||
15:40
vendethiel joined
|
|||
brrt | but that's a system build init | 15:43 | |
innit | |||
i mean shipping with the newest one | |||
flussence | that's gentoo's version of 1.7.5, which as far as I can tell from reading stuff, is mostly vanilla configure/make/make-install with no patches applied... | 15:47 | |
it passes "cc_cv_cflags__g=no" to configure but that's about it | |||
brrt | hmmkay | 15:48 | |
dalek | arVM/libuv-1.8: 7db3811 | brrt++ | / (2 files): Bump libuv to 1.8, change submodule url |
15:51 | |
JimmyZ | brrt: github.com/MoarVM/MoarVM/compare/l...6.1-update | 15:54 | |
brrt | JimmyZ: did that ever get merged? | 15:55 | |
because i kind of don't think that it did | |||
also | |||
we're at 1.8 by now | |||
JimmyZ | there was some patch on windows | ||
in that branch | 15:56 | ||
brrt | hmmmm | ||
lemmesee | |||
16:10
FROGGS joined
|
|||
dalek | arVM/libuv-1.8: 3c6c8fb | FROGGS++ | build/setup.pm: link to userenv.dll on windows for latest libuv |
16:19 | |
arVM/libuv-1.8: c2c796d | brrt++ | Configure.pl: Add submodule sync before update in Configure.pl This is needed to deal correctly with submodule repo url changes. |
|||
brrt | JiimmyZ++ | ||
16:25
pmurias joined
|
|||
pmurias | is there any randomness is the way ropes work on moar? | 16:25 | |
JimmyZ | pmurias: do you mean this one? github.com/MoarVM/MoarVM/tree/mast...rty/tinymt | 16:39 | |
16:41
zakharyas joined
|
|||
brrt dunno pmurias | 16:44 | ||
16:58
dalek joined
17:07
lucasb joined
17:36
ggoebel7 joined
18:01
Ven joined
18:20
domidumont joined
18:25
virtualsue joined
19:22
Ven joined
19:31
llfourn joined
19:39
brrt joined
20:05
virtualsue joined
20:32
llfourn joined
|
|||
brrt | libuv-1.8 gives me breakage in lowlevel locks and async sockets | 20:41 | |
let's try that on master hmm? | 20:42 | ||
20:57
FROGGS joined
|
|||
brrt | io socket async is still flappy/broken on 1.0 | 21:14 | |
21:31
llfourn joined
22:24
stmuk joined
22:28
lnx joined,
jnthn joined
22:32
llfourn joined
22:44
jnthn joined
22:45
stmuk joined
22:50
stmuk joined,
jnthn joined
22:51
lnx joined
23:28
pmurias joined
23:35
llfourn joined
|