Camelia, the Perl 6 bug

IRC log for #moarvm, 2013-08-13

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

All times shown according to UTC.

Time Nick Message
00:00 FROGGS joined #moarvm
00:13 lizmat joined #moarvm
00:55 benabik joined #moarvm
01:07 dalek MoarVM/libvu3: 172cc1c | jimmy++ | / (3 files):
01:07 dalek MoarVM/libvu3: updated oplist
01:07 dalek MoarVM/libvu3: review: https://github.com/MoarVM/MoarVM/commit/172cc1ccdc
01:13 colomon joined #moarvm
01:26 woolfy joined #moarvm
01:40 colomon joined #moarvm
01:41 dalek MoarVM: 28f5d3f | jimmy++ | src/io/procops.c:
01:41 dalek MoarVM: get acp at runtime when works on windows.
01:41 dalek MoarVM: review: https://github.com/MoarVM/MoarVM/commit/28f5d3f06d
01:42 JimmyZ BabsSeed, not_gerd: I just pushed my fix
01:42 JimmyZ in libuv3 branch
01:44 JimmyZ and now, getenvhash works well on CJK Windows
01:56 crab2313 joined #moarvm
02:23 FROGGS joined #moarvm
03:19 colomon joined #moarvm
03:22 FROGGS joined #moarvm
03:37 FROGGS joined #moarvm
03:37 crab2313 joined #moarvm
03:47 colomon joined #moarvm
04:00 FROGGS joined #moarvm
04:13 dalek joined #moarvm
04:44 birdwindupbird joined #moarvm
05:00 FROGGS joined #moarvm
05:59 JimmyZ hmm, looks like we don't need linenoise since libuv has readline support in src/*/tty.c
05:59 JimmyZ jnthn: ^^
06:00 JimmyZ but looks it support windows more better than linux
06:01 JimmyZ s/looks it/looks like it/
06:13 FROGGS joined #moarvm
06:14 FROGGS o/
06:31 BabsSeed o/
06:33 dalek MoarVM/libvu3: 62baae3 | jimmy++ | 3rdparty/libuv/ (5 files):
06:33 dalek MoarVM/libvu3: removed uv_fs_getstdin/uv_fs_getstdout/uv_fs_getstderr, libuv has tty support
06:33 dalek MoarVM/libvu3: review: https://github.com/MoarVM/MoarVM/commit/62baae329a
06:33 JimmyZ FROGGS: we may don't need linenoise, libuv is enough
06:34 FROGGS JimmyZ: yeah, I've read this in the logs
06:40 JimmyZ yeah, looks like libuv's tty is not supporting linux as well as windows
06:51 grondilu joined #moarvm
06:58 not_gerd joined #moarvm
06:58 not_gerd o/
06:59 not_gerd #46 *should* now work on darwin as well
07:02 JimmyZ not_gerd: linenoise may be excluded in #46
07:02 JimmyZ :)
07:02 JimmyZ I'm not sure, I'd like to ask jnthn again
07:05 not_gerd JimmyZ: will do so once readlineint_fh hass been migrated
07:05 BabsSeed Is there a todo or trac or something if I wanted to start coding on something? :P
07:05 not_gerd doesn't hurt to keep it there till then
07:07 JimmyZ BabsSeed: BabsSeed, replace apr by libuv :)
07:07 JimmyZ BabsSeed: you may want to ask jnthn++ ...
07:07 BabsSeed JimmyZ: Right now I'm still getting familiar with the code base :D
07:08 BabsSeed Was wondering if there's some kind of todo list that everyone grabs tasks from, or bug tracker
07:08 BabsSeed Back soon, commute time
07:08 JimmyZ BabsSeed: he could give you more answers :P
07:08 BabsSeed OK, will do
07:08 BabsSeed (or jnthn feel free to reply here :D)
07:08 BabsSeed o/
07:12 woolfy left #moarvm
07:13 crab2313 joined #moarvm
07:47 lizmat joined #moarvm
07:51 benabik joined #moarvm
08:00 JimmyZ hmm, I'm confused, Redis on windows uses both libuv and linenoise
08:31 lizmat joined #moarvm
08:40 lizmat joined #moarvm
08:41 segomos joined #moarvm
08:58 JimmyZ I think I figured how to integrate io part of libuv to MoarVM
08:59 JimmyZ *figured out
08:59 lizmat JimmyZ++
09:01 JimmyZ libuv support outputing "\033[41;37m" or something colored tty outputs as same as linux
09:02 JimmyZ jnthn: how do you think linenoise branch?
09:03 JimmyZ jnthn: and good morning
09:05 * not_gerd successfully built the nqp-cc for the first time
09:05 lizmat_ joined #moarvm
09:06 JimmyZ good
09:07 JimmyZ well, libuv support tty outputing well, not readline
09:07 jnthn JimmyZ: I've heard rather mixed things about linenoise. I want to do some research to figure out if it's what we want.
09:08 JimmyZ jnthn: ok, thanks.
09:29 not_gerd jnthn: diakopter: I need some feedback on https://github.com/MoarVM/MoarVM/pull/46
09:29 not_gerd it's not really urgent
09:30 not_gerd I just want to know if it's fine to do more workk on top of it
09:31 * not_gerd guesses one can always cherry-pick
09:34 diakopter not_gerd: what?
09:34 not_gerd diakopter: if been re.
09:34 diakopter not_gerd: you didn't see the dyncall branch?
09:34 not_gerd -*re-working the build system
09:35 diakopter where I also reworked the makefile completely
09:35 diakopter and integrated dyncall
09:35 not_gerd oO
09:35 JimmyZ diakopter: you didn't re-work Configure.pl :P
09:36 diakopter I didn't say I did
09:36 not_gerd so my work might not be totally useless ;)
09:36 diakopter well, if you had only asked ;)
09:36 JimmyZ pr/46 mostly re-work Congfigure part
09:40 not_gerd one of my goals was to get cross-compilation support going
09:41 not_gerd the RTEMS port of parrot had issues because the build system makes the unhealthy assumption that build==host
09:41 not_gerd I was trying to avoid that
09:41 not_gerd (also, I wanted to make my cagwin->mingw cross-setup work)
09:41 not_gerd *cygwin
09:42 diakopter JimmyZ & FROGGS & arnsholt & jnthn knew I'd integrated building dyncal
09:43 not_gerd diakopter: it's not really about dyncall, but the buildsystem refactor
09:43 JimmyZ yeah
09:44 diakopter okay; I didn't mention dyncall first; I mentioned reworking the makefile
09:45 diakopter I'm frustrated because I don't know what else I should've done to communicate it
09:46 JimmyZ I think makefile is almost the same, but some repeated working
09:46 not_gerd diakopter: well, I just started with trying to make my cross-setup work
09:47 not_gerd everything else naturally grew from that
09:47 diakopter right, it's simply a question of what should I do next
09:47 diakopter with that branch
09:47 diakopter I guess I'll delete it
09:48 not_gerd well, your branch couples to dyncall more tighly than mine
09:48 JimmyZ we can cherry-pick these commit
09:48 not_gerd I only do that for libuv right now - all other 3rdparty libs are basically pluggable
09:49 not_gerd ie even if we decided to go with my branch (that's of course not a foregone conclusion) your work might be salvagable
09:49 lizmat joined #moarvm
09:50 JimmyZ yeah
09:51 JimmyZ just wonders how jnthn thinks about #46 :P
09:51 diakopter no, we should definitely go with your branch; it's just I wish I knew you planned to do that; I wouldn't have wasted so much time learning how to use makefiles
09:52 lizmat_ joined #moarvm
09:52 not_gerd diakopter: well, Jul 20th I did not knwo I was going to do that ;)
09:52 JimmyZ If all is +1 to not_gerd++'s branch, I'd like cherry-pick main commits to master
09:53 JimmyZ and cherry-pick to other branches, ie> dyncall and libuv :-)
09:53 diakopter I think that's the wrong approach
09:53 not_gerd if there's a go-ahead, I can do some rebasing/squashing of my branch
09:53 diakopter pr#46 is fine
09:53 not_gerd ie prune all the false starts
09:54 not_gerd of course, one could also just make #46 the new master
09:54 not_gerd but before that, it needs some more testing
09:54 JimmyZ aye, I can test on both windows and linux, just Configure.pl && make :P
09:54 diakopter no, it'd be much more efficient to merge your branch and manually merge in the other things
09:55 JimmyZ and others args can be tested after merging
09:56 diakopter jnthn: do you have a comment?
10:01 * not_gerd needs to go buy groceries or he won't eat today
10:13 diakopter I think it should be merged, and if jnthn decides against linenoise, simply remove linenoise
10:40 Util joined #moarvm
10:44 * not_gerd bought groceries, starts cooking
11:12 lizmat joined #moarvm
11:20 crab2313 joined #moarvm
11:28 JimmyZ Good evening
11:31 FROGGS hi JimmyZ
11:32 colomon joined #moarvm
11:33 JimmyZ hello FROGGS
11:49 JimmyZ http://www.corsix.org/content/first-dynasm-example I just find this, easy to understand
11:50 JimmyZ and small
11:53 lizmat_ joined #moarvm
12:00 lizmat joined #moarvm
12:07 lizmat joined #moarvm
12:33 crab2313 joined #moarvm
13:11 cognominal joined #moarvm
13:47 lizmat joined #moarvm
14:42 BabsSeed Afternoon all
14:49 dalek MoarVM/libuv2: 0aa87fb | jimmy++ | src/6model/reprs/MVMOSHandle.h:
14:49 dalek MoarVM/libuv2: starting get uv_handle_t into MVMOSHandle
14:49 dalek MoarVM/libuv2: review: https://github.com/MoarVM/MoarVM/commit/0aa87fbc27
14:49 JimmyZ BabsSeed: good afternoon
14:52 masak BabsSeed: $afternoon.good()
14:55 colomon joined #moarvm
15:52 FROGGS joined #moarvm
17:53 benabik FWIW I'd rather keep history cleaner and do the cherry-pick/rebasing dance so we only put the master-related bits on master and let the stuff for branches merge with the branches.
17:54 * benabik fetches and tests pr46
17:57 benabik I'd call that something less than awesome...  Piles of warnings and then "ld: warning: ignoring file 3rdparty/libuv/libuv.a, file was built for archive which is not the architecture being linked (x86_64): 3rdparty/libuv/libuv.a"
18:00 not_gerd joined #moarvm
18:00 not_gerd o/
18:00 benabik I'll put up a gist with the build output in a moment.  Just did `make clean` to make sure I get all of it.
18:00 not_gerd benabik: os/arch/compiler?
18:00 benabik not_gerd: OS X / x86_64 / clang
18:01 not_gerd benabik: the warnings come from -Weverything, my default mode of developing in C
18:01 benabik Huh.  `make clean && perl Configure.pl && make` gets me "No rule to make target `3rdparty/libuv/src/unix/darwin.c', needed by `3rdparty/libuv/libuv.a'.  Stop."
18:01 not_gerd I did not yet put in appropriate -Wno-* options
18:02 benabik not_gerd: I love -Wall, but the goal of using it is so the code is clean of warnings, not so that I get kilobytes of warning output when I build.  :-/
18:03 not_gerd benabik: moarvm is not yet 'released' - fix the warnings or add a -Wno-* to Configure.pl if that's impossible
18:03 not_gerd I just didn't get around to that yet
18:03 not_gerd eventually, the build should be warning free
18:04 benabik not_gerd: Disable -Weverything until you have the warnings right?  Having it spew that much text actually slows the build as the terminal emulator has to keep up.
18:04 benabik Huh.  `make clean` seems to be removing source files.
18:05 not_gerd that... shouldn't be happening
18:05 not_gerd could you reset the repository, run configure and gist your Makefile?
18:07 benabik https://gist.github.com/Benabik/6223922
18:07 not_gerd thanks
18:08 benabik Removes 3rdparty/libuv/src/unix/{darwin-proctitle.c,darw​in.c,fsevents.c,internal.h,kqueue.c,proctitle.c}
18:09 not_gerd yes, I forgot to replace .c by @obj@ in build/Makefile.in
18:09 segomos not_gerd: i was working on fixing it for darwin
18:10 segomos wasn't sure if you had already got it working
18:10 not_gerd segomos: not yet, but soon ;)
18:11 segomos where did all of the 'use' statements go that were in configure?
18:12 not_gerd benabik: pushed 2 commits, couls´d you try again?
18:12 not_gerd segomos: in my branch?
18:12 segomos not_gerd: yea
18:12 segomos pr/46
18:12 not_gerd that's a completely new Configure.pl
18:12 segomos not_gerd: i see
18:13 benabik not_gerd: `make clean` still removes `3rdparty/libuv/src/unix/internal.h`
18:14 not_gerd benabik: fixed
18:14 not_gerd segomos: I tried to hit the right balance between not flexible enough/too complicated
18:14 not_gerd not guarantees that I got it right
18:15 not_gerd after all, it's a build system - they suck almost by definition
18:16 benabik not_gerd: Huh.  Now not only does `make clean` not remove tracked files, it's not complaining about the wrong arch on libuv.a
18:17 not_gerd well, putting source files into an archive can lead to unexpected results ;)
18:17 benabik Ah.
18:18 not_gerd benabik: it links correctly now?
18:19 benabik not_gerd: Seems to?
18:19 not_gerd well, does the moarvm executable run or not?
18:19 benabik not_gerd: Well it gives me the "no input file" error.
18:20 not_gerd that's the correct behaviour if you did not provide an input file ;)
18:20 not_gerd benabik++
18:20 not_gerd thanks for the help
18:20 * benabik needs to update his NQP to test.
18:22 colomon joined #moarvm
18:29 segomos not_gerd: not sure i did this right, i added build instructions to %UV hash for libuv and i'm building moarvm successfully on osx, 2 q's
18:29 segomos a) is that proper place for build
18:29 segomos b) want me to push if a is yes
18:30 not_gerd segomos: did you see the commits I did?
18:30 not_gerd they made it build for benabik
18:31 segomos yea i pulled it in and still got libuv built for wrong version..let me copy this file somewhere and re-pull/build
18:31 not_gerd os-specific build instructions go to %OSNAME => { -thirdparty => { ... } }
18:32 not_gerd for more involved stuff, you can add code after the defaults detection
18:32 not_gerd it's done this way to detect x64 on windows
18:33 segomos maybe i'm not doing something right but i get the same 'ld: warning: ignoring ... architecture different' error
18:33 segomos s/error/warning/
18:34 not_gerd segomos: did you pull, reset and clean the branch?
18:34 not_gerd if it still fails, could you gist the generated Makefile?
18:35 benabik segomos: Probably didn't reconfigure, or didn't grab the last commit.
18:43 segomos i did perl Configure.pl && make
18:43 benabik segomos: Try `make clean ; git status`.  If there are any files missing, you probably don't have the latest commit.
18:46 segomos make: *** No rule to make target `3rdparty/libuv/src/unix/darwin.c', needed by `3rdparty/libuv/libuv.a'.  Stop.
18:46 benabik Yeah, missing a few commits.
18:52 segomos this is pull request 46 still?
18:55 benabik Yes.
18:55 not_gerd segomos: assuming you're still on the correct branch, do a `git reset --hard HEAD && git pull`
18:56 not_gerd you should git stash if you have changes you want to keep
18:56 * not_gerd probably should have started with that ;)
18:57 segomos that's what i've been trying, maybe i'm not checking out pr/46 properly - git checkout pr/46
18:57 segomos is what i've been doing
18:58 segomos git fetch origin grabbed pr/46
18:59 PerlJam segomos: what version of git are you using?
18:59 segomos 1.7.12.4
18:59 benabik `git checkout pr/46` will checkout your own local version if you've done it before.
18:59 benabik Once you're on your pr/46, you need to pull
18:59 segomos i have no upstream tracking ..
19:00 PerlJam I'm not sure if that worked in 1.7
19:00 benabik or `git merge origin/pr/46`
19:00 segomos downloading 1.8
19:00 PerlJam segomos: git checkout -b pr/46 -t origin/pr/46   # while verbose, should work.
19:03 benabik `git checkout foo` learned to checkout origin/foo in 1.6.6.
19:23 segomos that worked, thanks
19:23 segomos builds now too :)
19:23 not_gerd \o/
19:30 segomos commands for git 1.8 are slightly different :(
19:39 crab2313_ joined #moarvm
19:59 colomon joined #moarvm
20:00 crab2313 joined #moarvm
20:12 not_gerd bye, #moarvm
20:12 not_gerd left #moarvm
20:13 Tene joined #moarvm
20:33 benabik joined #moarvm
20:42 lizmat joined #moarvm
21:23 lizmat joined #moarvm
21:46 colomon joined #moarvm
21:50 benabik joined #moarvm
22:46 woolfy joined #moarvm
23:28 benabik joined #moarvm

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