|
00:55
tokuhiro_ joined
03:50
vendethiel joined
07:18
FROGGS joined
07:38
brrt joined
|
|||
| brrt | good UGT | 07:39 | |
| nwc10 | good UGT, brrt | 07:41 | |
| brrt | hey | 07:42 | |
| you guys fixed it? :-o | 07:43 | ||
| jnthn++ | |||
| oh, the 6th arg explains why it didn't go wrong for jnthn | 07:44 | ||
| 6th win32 arg is stack-passed, on unix it's register-passed | 07:45 | ||
| FROGGS: MVM_JIT_REG_VAL and MVM_JIT_REG_VAL_F are really the same op. MVM_JIT_LITERAL is meant for smallish integers, MVM_JIT_LITERAL_64 and friends not | 07:47 | ||
| jnthn++ | 07:48 | ||
| we can actually replace that, i think, with a fits_in_32_bit check | 07:51 | ||
| FROGGS | :o) | 08:10 | |
|
08:20
vendethiel joined
|
|||
| brrt | anyway, thanks for fixing that :-) | 08:27 | |
| brrt can, somewhat uselessly, confirm that it works for him too | 09:06 | ||
| FROGGS | hehe, that's still nice to know :o) | 09:07 | |
| brrt: btw, what are your next steps? fix more bugs? | |||
| brrt | ehm, yes, i guess so | 09:09 | |
| there are still vast maps of NYI in the JIT compiler that should be implemented | |||
| but first and foremost i want to remove bugs | |||
| there is a bug somewhere in the new JIT that seems to return the wrong value in a routine | 09:10 | ||
| i'm going to see if i can golf that | |||
| and why it's broken | |||
| FROGGS | brrt: I can't help you in any area, right? | 09:11 | |
| brrt | hmmm | ||
| well, how good are you with GDB? :-P | |||
| i can tell you what i know is going wrong, and maybe you have an idea | |||
| actually, the truth is, i know very little of what is going wrong yet | 09:12 | ||
| FROGGS | well, fell free to paste a test case | ||
| brrt | hmm | 09:13 | |
| FROGGS | feel* | ||
| brrt | my 'testcase' is basically 'run moar in gdb. see where it blows up' | ||
| FROGGS | how do you run moar? nqp-m -e'' ? | 09:14 | |
| brrt | gist.github.com/bdw/56dd8b724f55e79183ce | 09:15 | |
| is what i do | |||
| MVM_JIT_EXPR_ENABLE is what is necessary to enable the expression compiler | |||
| FROGGS | looks like it blows up here in the elems op | 09:23 | |
| brrt | aye | 09:24 | |
| but why? | |||
| i'd have to look at the code that is being compiled | |||
| FROGGS | valgrind is unhappy too | ||
| brrt | oh, that helps | ||
| about what? | |||
| FROGGS | gist.githubusercontent.com/FROGGS/...tfile1.txt | 09:25 | |
| brrt | aw, that's awful | 09:27 | |
| FROGGS | it is | 09:29 | |
| brrt | hmm | ||
| i wonder why | |||
| FROGGS++ that's a huge help though | 09:42 | ||
| (my suspicion is that valgrind makes it crash because it zeroes out memory?) | |||
| oh, hmm | 09:50 | ||
| you may want to do a make reconfig before you run even-moar-jit? | |||
| that's basically the generic advice in any case :-) | |||
| FROGGS | I recompiled moar, yes | 09:51 | |
| brrt | i'm not getting a crash, though | ||
| but still | |||
| FROGGS | not nqp though, but I think that's no needed | ||
| hmmm | |||
| one of the Invalid read of size 8 seems to be about info->tile->right_sym at tile.c:197 | 10:01 | ||
| brrt | aye | 10:03 | |
| i see why that should be so | |||
| FROGGS | do tell | 10:05 | |
| brrt | it's realloced in assign_tile when we need more space due to tile resolution | 10:07 | |
| however | |||
| when i remove that | 10:08 | ||
| it still hapens | |||
| happens | |||
| dalek | arVM/even-moar-jit: c5733aa | brrt++ | src/jit/ (2 files): Don't cache from realloced array This silences valgrind, but does not solve the wrong-value-return bug |
10:12 | |
| FROGGS | yeah, the only oddness thath stays is this: | 10:15 | |
| ==10392== Jump to the invalid address stated on the next line | |||
| ==10392== at 0x0: ??? | |||
| ==10392== by 0x4F9B2EA: MVM_interp_run (interp.c:2603) | |||
| brrt | hmmmm | 10:17 | |
| FROGGS | what happens when you do that? memcpy(88582584, 88582584, 40) | 10:20 | |
| ohh wait | |||
| nvm, that was my stupidity | 10:21 | ||
| brrt | where should i have done it? :-) | 10:23 | |
| FROGGS | no havn't :o) | ||
| lunch & | 10:24 | ||
| brrt also lunch & | 10:25 | ||
|
10:50
vendethiel joined
10:55
colomon joined
11:17
brrt joined
11:52
FROGGS joined
12:01
brrt joined
|
|||
| brrt | ok, i'm now diffing the code pieces | 12:24 | |
| FROGGS | code pieces of what? | 12:28 | |
| generated code? | |||
| old jit vs new jit? | |||
| brrt | aye | 12:32 | |
| something seems to be pretty wrong... | 13:05 | ||
| [Coke] | new jit, same as the old jit. | 13:10 | |
|
13:10
rarara joined
|
|||
| brrt | new jit, worse than the old jit :-( | 13:10 | |
| [Coke] | sorry, that was a reference to the band the Who, not your coding skills. :) | 13:11 | |
|
13:13
Peter__R joined
|
|||
| brrt | :-) | 13:16 | |
| i wonder what the JIT thought it was doing | 13:21 | ||
| FROGGS | O.o | ||
|
13:29
colomon joined
|
|||
| brrt | there is a call in the middle of that code | 13:31 | |
| and it's weird | 13:32 | ||
| hmmm | 13:38 | ||
| maybe it's the write barier i'm looking at | |||
|
14:12
FROGGS joined
|
|||
| FROGGS | o/ | 14:15 | |
|
15:08
synbot6 joined
|
|||
| timotimo | o/ | 15:43 | |
|
16:00
Peter_R joined
16:24
vendethiel joined
16:28
Peter_R joined
17:12
Peter_R joined
17:31
Peter_R joined
17:38
Peter_R joined
17:54
Peter_R joined
18:00
Peter_R joined
19:27
zakharyas joined
19:41
FROGGS joined
22:57
Peter_R joined
|
|||