|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:12
lizmat joined,
p6bannerbot sets mode: +v lizmat
00:16
lizmat left
|
|||
| MasterDuke | timotimo: any idea why gist.github.com/MasterDuke17/e82d3...769fedf684 would cause `Argument "Expected 3 operands and 0 params, got 0" isn't numeric in addition (+) at tools/expr-template-compiler.pl line 257, <STDIN> line 2399.` ? | 00:37 | |
| timotimo | sounds like an error not being properly output'd | 00:38 | |
| not sure what's wrong in the templates | 00:40 | ||
| MasterDuke | hm, if i remove the macro i get `Mismatched type, got reg expected voidfor do (decodertakebytes ARRAY(0x5600ac023b60)) at tools/expr-template-compiler.pl line 286, <STDIN> line 2398.` | 00:42 | |
| Geth | MoarVM: cb89b74160 | MasterDuke17++ (committed using GitHub Web editor) | tools/expr-template-compiler.pl Add missing whitespace to die output |
00:45 | |
| MasterDuke | timotimo: the `Expected 3 operands and 0 params, got 2` comes from the macro call | 00:59 | |
| timotimo | why, though? | 01:00 | |
| MasterDuke | i thought maybe it was the string, so i changed it to (const_ptr "foo") and same error | ||
| and same with just (tc) | |||
| i'm trying changes in the macro itself, but nothing i do changes the error | 01:03 | ||
| timotimo | i'd have to defer to brrt :( | 01:04 | |
| m: srand(10**1000) | 01:07 | ||
| camelia | Cannot unbox 3322 bit wide bigint into native integer in block <unit> at <tmp> line 1 |
||
| timotimo | mhm, kind of makes sense | ||
| m: int64.pick | |||
| camelia | ( no output ) | ||
| timotimo | m: int64.pick.say | ||
| camelia | (int64) | ||
| timotimo | haha | ||
| m: int64.Range.pick.say | |||
| camelia | -5273172063679231472 | ||
| timotimo | m: srand(-1) | ||
| camelia | ( no output ) | ||
| timotimo | OK | 01:08 | |
| i'm going to bed | 01:09 | ||
| good luck! | |||
| MasterDuke | thanks. later... | ||
|
01:18
Redfoxmoon left
|
|||
| MasterDuke | ah, got it compiling. now does it run... | 01:23 | |
|
01:24
Redfoxmoon joined,
p6bannerbot sets mode: +v Redfoxmoon
01:45
drot8 joined
01:48
drot8 left
02:04
dalek left
02:05
dalek joined,
p6bannerbot sets mode: +v dalek
02:45
dalek left
02:54
dalek joined
02:55
p6bannerbot sets mode: +v dalek
03:25
camelia left
03:27
nine left
03:32
MasterDuke left,
camelia joined
03:33
nine joined,
p6bannerbot sets mode: +v camelia
03:34
p6bannerbot sets mode: +v nine
03:55
lizmat joined,
p6bannerbot sets mode: +v lizmat
03:59
lizmat left
|
|||
| Geth | MoarVM: MasterDuke17++ created pull request #953: Add a macro for decoder_ensure_decoder and ... |
04:44 | |
| MoarVM: e0eed2b188 | MasterDuke17++ (committed using GitHub Web editor) | tools/expr-template-compiler.pl Add operator to die message and fix error about... the message not being numeric. |
04:53 | ||
|
05:07
Redfoxmoon left,
Redfoxmoon joined,
karatkievich.freenode.net sets mode: +v Redfoxmoon,
p6bannerbot sets mode: +v Redfoxmoon
06:09
domidumont joined
06:10
p6bannerbot sets mode: +v domidumont
06:14
AlexDaniel left
|
|||
| Geth | MoarVM: 62946ae642 | (Nick Logan)++ | 2 files Fix sprintf formatting for the debugserver port Fixes clang warnings such as "format specifies type 'unsigned long' but the argument has type 'MVMuint64' (aka 'unsigned long long')" See: travis-ci.org/rakudo/rakudo/jobs/424479274#L228 |
06:14 | |
| MoarVM: c114c1e67a | niner++ (committed using GitHub Web editor) | 2 files Merge pull request #952 from ugexe/fix-clang-warnings Fix sprintf formatting for the debugserver port |
|||
|
06:19
domidumont left
07:51
ilogger2 joined
07:52
p6bannerbot sets mode: +v ilogger2
08:09
robertle joined,
p6bannerbot sets mode: +v robertle
08:10
brrt joined,
p6bannerbot sets mode: +v brrt
|
|||
| brrt | \o | 08:11 | |
| brrt-to-the-future.blogspot.com/201...pdate.html | 08:12 | ||
| i accidentally nuked my magit installation | 08:15 | ||
| and now I'm stuck with the old one, and I hate it | 08:16 | ||
| tadzik | brrt: "harmless at runtime, it is nevertheless confusing, so the template compiler" -- I think you accidentally a few words :) | 08:18 | |
| brrt | tadzik: thanks | 08:20 | |
|
08:20
Kaypie joined
08:21
p6bannerbot sets mode: +v Kaypie
08:22
lizmat joined,
p6bannerbot sets mode: +v lizmat
08:24
brrt left
|
|||
| Geth | MoarVM/jit-expr-optimizer: 4 commits pushed by (Bart Wiegmans)++ | 08:25 | |
|
08:26
brrt joined
08:27
p6bannerbot sets mode: +v brrt
|
|||
| lizmat | brrt: "it is nevertheless confusing, so the template compiler " seems the end of the sentence is missing | 08:29 | |
| brrt++ # brrt-to-the-future.blogspot.com/201...pdate.html | 08:30 | ||
| brrt | thank you | 08:31 | |
| fixed it | 08:32 | ||
| Geth | MoarVM/fork-safety: ec8bd3d909 | (Bart Wiegmans)++ | 5 files [Spesh] Tear down spesh worker in full_cleanup If the spesh worker is still active the GC can try to free a lock that is being held, which leads to an abort in libuv. Stopping the thread makes full-cleanup work (although it still leaks). Thread is stopped by signalling with VMNull. Ideally we'd unshift this to the front of the queue, but that is more complicated. |
08:33 | |
| MoarVM/fork-safety: 5e1e301402 | (Bart Wiegmans)++ | 3 files [ConcBlockingQueue] Unmoving body The locks structure was already malloc'd, and I extended this to include the entire body. This prevents the GC from moving the body while we are blocked on acquiring a lock, simplifying the code, and has negligible cost. |
|||
| brrt | ^rebase | ||
| MoarVM/fork-safety: fe8b30ec79 | (Bart Wiegmans)++ | 3 files [IO] Stop event loop thread Stop loop with uv_stop(uv_loop) and join the thread |
|||
|
09:32
zakharyas joined
09:33
p6bannerbot sets mode: +v zakharyas
09:45
brrt left
09:47
domidumont joined,
p6bannerbot sets mode: +v domidumont
09:50
AlexDaniel joined,
p6bannerbot sets mode: +v AlexDaniel
10:08
domidumont left
11:07
zakharyas left,
zakharyas joined
11:08
p6bannerbot sets mode: +v zakharyas
11:23
zakharyas left
11:42
brrt joined,
p6bannerbot sets mode: +v brrt
12:33
ggoebel joined
12:34
p6bannerbot sets mode: +v ggoebel
|
|||
| timotimo | brrt: hack.p6c.org/~timo/perl6_log.html | 12:47 | |
| brrt: hack.p6c.org/~timo/moarvm_log.html | 12:48 | ||
| brrt | timotimo++ | 12:53 | |
|
13:05
zakharyas joined
13:06
p6bannerbot sets mode: +v zakharyas
|
|||
| brrt | I've decided to ignore the debugserver thread for the purposes of fork() | 13:25 | |
| and... | |||
| (as in, if there is a debugserver thread, we're not going to do fork(), atm) | 13:26 | ||
|
13:27
AlexDaniel left
|
|||
| timotimo | fwiw, the debug server client is supposed to already cope with disconnection of the server ... or at least i seem to remember that | 13:33 | |
| brrt | that's fine; the problem is I can't *stop* the debug server thread :-) | 13:49 | |
| becuase there is no safe, portable way to interrupt a thread waiting on accept() | |||
| or at least, I don't htink there is | |||
| timotimo | oh, i remember that | 13:52 | |
| hm, so how does libuv implement accepting listen sockets? | 13:53 | ||
| they can't "only accept if there's a socket trying to connect", or can they? | |||
| geekosaur | if it's O_NONBLOCK accept() would return EAGAIN if nothing's in the listen queue | 14:06 | |
|
14:07
lizmat left
|
|||
| brrt | yeah, but it is blocking | 14:07 | |
| libuv can do accept(), yes | |||
| but I don't want to have to rewrite the debugserver today :-) | 14:08 | ||
| timotimo | um, why don't we just connect to the socket locally to get out of the accept? | 14:13 | |
|
14:39
brrt left
|
|||
| jnthn | brrt++ # blog | 15:12 | |
|
15:27
Kaypie is now known as Kaiepi
15:32
AlexDaniel joined,
p6bannerbot sets mode: +v AlexDaniel
15:47
robertle left
16:06
brrt joined
16:07
p6bannerbot sets mode: +v brrt
|
|||
| brrt | timotimo: that is insane, but not so bad actually | 16:07 | |
|
16:48
robertle joined,
p6bannerbot sets mode: +v robertle
16:49
fake_space_whale joined,
Kaiepi left,
p6bannerbot sets mode: +v fake_space_whale
16:57
Kaiepi joined
16:58
p6bannerbot sets mode: +v Kaiepi
|
|||
| brrt | I'd have to publish the socket though | 17:05 | |
|
17:14
lizmat joined
17:15
p6bannerbot sets mode: +v lizmat
|
|||
| timotimo | or just publish a "interrupt the server" function | 17:19 | |
| brrt | oh, yeah, or that | ||
| :-) | |||
|
17:31
brrt left
18:18
zakharyas left
18:49
domidumont joined,
p6bannerbot sets mode: +v domidumont
18:54
domidumont left
|
|||
| Geth | MoarVM: jstuder-gh++ created pull request #954: Implement 'cast_load_addr' tile for exprjit |
20:18 | |
|
21:05
MasterDuke joined,
p6bannerbot sets mode: +v MasterDuke
21:07
MasterDuke left,
MasterDuke joined,
herbert.freenode.net sets mode: +v MasterDuke,
p6bannerbot sets mode: +v MasterDuke
|
|||
| samcv | good * | 21:07 | |
| timotimo | hello samcv | 21:12 | |
| samcv | gonna fix this bug right now github.com/MoarVM/MoarVM/issues/924 | 21:39 | |
| lizmat | ++samcv | 21:42 | |
| samcv | essentially going to take it out of the database since there are points that don't exist that have this property. unlike every single other property | 21:43 | |
| so we can't be sure it's in the database | |||
| i.e. while others default to default property values if they aren't specified, this one does | |||
| err does not | |||
|
22:26
robertle left
|
|||
| timotimo | oooh | 22:27 | |
| this program bails out because trunc_i8 is missing from the jit! | 22:28 | ||
| 33.47user 0.13system 0:33.30elapsed 100%CPU (0avgtext+0avgdata 175504maxresident)k | 22:33 | ||
| 45.69user 0.17system 0:45.70elapsed 100%CPU (0avgtext+0avgdata 180612maxresident)k | |||
| the top one is with jit (because i'm not using int8 for that one variable) and the bottom one is with int8 there | 22:34 | ||
| MasterDuke | !! | 22:55 | |