|
Parrot 3.10.0 "Apple Pi" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC Set by moderator on 22 November 2011. |
|||
|
00:48
whiteknight joined
|
|||
| whiteknight | good evening, #parrot | 00:53 | |
| msg NotFound I'm trying to run that fly.winxed example, but I don't have opengl_defines.pasm. Where do I get that file? | 01:06 | ||
| aloha | OK. I'll deliver the message. | ||
|
01:26
alvis joined
|
|||
| dalek | sella: d23f71d | Whiteknight++ | src/unstable/random/MersenneTwister.winxed: Convert constants in MersenneTwister from attributes to constant ints. |
01:43 | |
| sella: 61e6c3b | Whiteknight++ | s (2 files): Add in a Rosella.Random namespace with a method to get a default random number generator. |
|||
| sella: 5a540ef | Whiteknight++ | / (7 files): Add in a Box-Muller transformation to turn the Marsenne twister into a normal distribution generator. Add basic histograms to the example program |
|||
| sella: 9635fe2 | Whiteknight++ | / (3 files): fix the interface for the normal number generator to be the same as the uniform number generator |
01:50 | ||
|
04:01
jsut joined
|
|||
| dalek | rrot: bbe82b9 | NotFound++ | src/pmc/structview.pmc: refactor repeated throws into static functions |
05:20 | |
|
05:42
rfw joined
05:50
benabik joined
|
|||
| NotFound | msg whiteknight the opengl includes are created during the build if you have opengl configured | 05:58 | |
| aloha | OK. I'll deliver the message. | ||
| benabik | o/ #parrot | 06:01 | |
| torrap# \\o | 06:06 | ||
|
08:26
lucian joined
08:44
mj41 joined
09:21
lucian joined
12:13
Psyche^ joined
12:27
whiteknight joined
|
|||
| whiteknight | good morning, #parrot | 12:27 | |
| msg NotFound: error while building rosella: Method 'emit' not found for invocant of class 'Winxed;Compiler;FunctionId' | 12:56 | ||
| aloha | OK. I'll deliver the message. | ||
| whiteknight | msg NotFound the error happens with the current winxed snapshot and master | 13:05 | |
| aloha | OK. I'll deliver the message. | ||
|
13:19
bluescreen joined
13:52
benabik joined
|
|||
| dalek | sella: d19b688 | Whiteknight++ | src/query/ (4 files): Workaround for a winxed bug with the default sort comparer. Fix the Array Provider .foreach and .map methods to iterate in the forward direction as expected |
13:59 | |
|
14:23
whiteknight joined
14:36
benabik joined
14:38
PacoLinux joined
|
|||
| dalek | nxed: 0a49b0c | NotFound++ | winxedst1.winxed: ooops, FunctionId lacked the emit method. whiteknight++ |
14:38 | |
| whiteknight | NotFound++ Quick Fix! | ||
| NotFound: When is the inline branch going to merge? I have lots of places I want to use it in Rosella | 14:40 | ||
|
14:40
cosimo_ joined
|
|||
| NotFound | whiteknight: I found some problems and decided that I'm going to first add an stage 2 and then integrate inline into it rather than in stage 1. I'm doing that in a local branch. | 14:42 | |
| whiteknight | oh, what's the problem in stage 1? | ||
| NotFound | Adding thing to stage 1 implies dealing with the limitations and bad diagnostics of stage 0. | 14:43 | |
| whiteknight | so stage2 is going to diverge from the stage1 implementation? | 14:45 | |
| NotFound | whiteknight: yeah. | ||
| There will be stage 1 and stage 2 winxed sources, and stage 3 will be stage 2 compiled with itself. | 14:46 | ||
|
14:47
lucian joined
|
|||
| dalek | rrot: 9279f64 | mls++ | compilers/pct/src/PCT/Node.pir: PCT/Node.pir: use 'append' instead of creating an iterator and pushing every element |
16:03 | |
| Heuristic branch merge: pushed 26 commits to winxed/stage2_inline by NotFound | 17:00 | ||
| nxed/stage2_inline: bedc804 | NotFound++ | winxedst1.winxed: fix merge master conflict |
17:06 | ||
| nxed/stage2_inline: 58b22fd | NotFound++ | winxedst2.winxed: diagnose inline with local functions not allowed |
17:15 | ||
| nxed/stage2_inline: a669e35 | NotFound++ | winxedst2.winxed: generic diagnostic for inline not allowed |
17:22 | ||
| nxed/stage2_inline: d431d64 | NotFound++ | winxedst2.winxed: diagnose yield inline |
17:24 | ||
| nxed/stage2_inline: c45da47 | NotFound++ | winxedst2.winxed: implement clone is some more operators |
17:34 | ||
| nxed/stage2_inline: 2c11441 | NotFound++ | examples/fly.winxed: use inline in example fly |
17:39 | ||
| rrot: d4331ed | nine++ | t/pmc/task.t: Remove timing assumptions from t/pmc/task.t |
17:45 | ||
|
17:51
fperrad joined
|
|||
| dukeleto | ~~ | 18:21 | |
|
18:45
zby_home joined
|
|||
| whiteknight | hello duke | 18:56 | |
| dalek | nxed/stage2_inline: 64dbc7e | NotFound++ | winxedst2.winxed: implement clone in MultiStatement and pirops |
18:59 | |
| dukeleto | whiteknight: how was your turkey day? | 19:00 | |
| whiteknight | dukeleto: very good, thanks. My grandmother's 80th birthday party is happening tomorrow night, so family from all over the country has flown in for the occasion | 19:03 | |
| so the holiday yesterday was more heavily populated than usual | |||
| how was your day? | 19:05 | ||
| dukeleto | whiteknight: wow! sound awesome. I am at a large family gathering in Olympia, Washington, which we are calling "Occupy Olympia" | ||
| because about 20 people are staying/camping in one house | 19:06 | ||
| whiteknight | heh, "Occupy" is t3h memez | ||
| dukeleto | or "occu-pie" | 19:09 | |
| dukeleto is occupying some Pecan Pie | |||
| whiteknight: did you happen to keep the gci tasks you were making in a git repo somewhere? | 19:10 | ||
| whiteknight: i would like to tweak some of your tasks to make new tasks | |||
| whiteknight | dukeleto: No, I have not. I've been dumping them from my head into melange | ||
| They have a csv export functionality | |||
| and I've basically be copying your template | |||
| dukeleto | whiteknight: ok | 19:19 | |
| whiteknight: i like putting the tasks in a git repo so we can refer to them in future years | 19:20 | ||
| whiteknight | I can try to do some of that | ||
| dukeleto | whiteknight: i am concerned that many of the tasks are submitting code with no tests | ||
| whiteknight: so i think we should make some kind of stipulation for future tasks that there be some kind of basic test to prove that stuff works | |||
| whiteknight | I've been asking for example code for most of the rosella stuff, but I mostly haven't been asking for automated tests, for them. | 19:21 | |
| For stuff in the Parrot repo, I do ask for tests, or at least to run existing tests to prove no breakage | |||
| dukeleto | whiteknight: sure. For example, the mersenne twister stuff needs a few tests and I was thinking that would make a good task | 19:22 | |
| whiteknight: we can turn the need for tests into more tasks | |||
| whiteknight | yes, that would. I needed to do a little work turning that into a proper library first, because Rosella's test suite is completely oriented towards libraries | 19:23 | |
| I added a Box-Muller implementation this morning too, using the twister as the data source, so that's going to need tests | |||
| dukeleto | whiteknight: this might help : github.com/leto/math--primality/tree/master/t | 19:27 | |
| whiteknight: and github.com/leto/math--primality/tree/master/xt | |||
| whiteknight: lots of good test data that can be stolen | 19:28 | ||
|
19:28
jsut_ joined
|
|||
| whiteknight | the problem with a PRNG is how do you actually test it in a way that the test never gives a false negative? | 19:28 | |
| I don't want to be getting a statistically certain number of false bug reports | |||
| Or we can write tests that are so loose and forgiving that they don't tell us anything about the generator | 19:29 | ||
| dukeleto | whiteknight: no no no, people have solved that problem already :) | ||
| whiteknight: en.wikipedia.org/wiki/Diehard_tests | 19:30 | ||
| whiteknight | Most of those tests I see would still allow for false negatives | 19:32 | |
| I'll keep that link handy, for now, and I will accept submissions for tests using that | 19:33 | ||
| dukeleto | whiteknight: cool. Basically, if you run statistical tests with the right number of iterations, false negatives can be made very rare | 19:34 | |
| whiteknight: because you get to define how much tests need to deviate from "what you expect" to fail | 19:35 | ||
| whiteknight | right, but "very rare" and "not possible" are still different things. | ||
| dukeleto | whiteknight: indubiously | ||
|
19:35
schmooster joined
|
|||
| dukeleto | Every time I bang my fist on the kitchen table, there is a non-zero chance that all the atoms in my hand will miss all the atoms in the table. | 19:35 | |
| whiteknight | As far as I am concerned, at least for now, so long as the numbers plausibly appear to be random to the untrained human eye, I'm satisfied | ||
| eventually I'll want a higher standard, but I don't need it now | 19:36 | ||
| dukeleto | whiteknight: works for me, for now | ||
| whiteknight | (again, I will always accept code submissions, I'm just talking about budgeting my own time) | ||
| dukeleto | whiteknight: yeps. | ||
| dukeleto goes for a post-pie walk | |||
| whiteknight | have fun | ||
| dalek | nxed/stage2_inline: b79afdc | NotFound++ | winxedst2.winxed: implement clone in do, while, break and continue |
19:37 | |
| sella: 108ac81 | Whiteknight++ | src/unstable/random/ (2 files): Implement BoxMullerNormal.get_int(). It's ugly but it works. Small optimization to MersenneTwister, use integers for mti calculations. |
19:39 | ||
| sella: cc7a6ae | Whiteknight++ | src/unstable/random/BoxMullerNormal.winxed: Completely calculate the spare value in BoxMuller immediately, since the center and stddev are invariants |
|||
|
19:40
mj41 joined
|
|||
| dalek | nxed/stage2_inline: 300598a | NotFound++ | winxedst2.winxed: style fixes |
20:01 | |
| sella: 6b0ea74 | Whiteknight++ | src/include/Random.winxed: +An include file for Random |
20:02 | ||
| sella: 7d8585d | Whiteknight++ | s (5 files): Move the knuth-fisher-yates shuffle implementation to the Random library. Update it to use the twister. Make Random a dependency of Query |
20:07 | ||
|
20:10
rfw joined
20:48
alvis joined
|
|||
| dalek | nxed/stage2_inline: 38dcd26 | NotFound++ | winxedst (2 files): fix a problem with local functions sharing used subids with its outers |
20:54 | |
| nxed/stage2_inline: 5e03a27 | NotFound++ | winxedst2.winxed: use some more binders |
21:23 | ||
| nxed: 3d29f54 | NotFound++ | winxedst (2 files): FunctionId.emit |
21:28 | ||
| nxed: 99b4e0b | NotFound++ | winxedst1.winxed: fake inline in stage 1 |
|||
| nxed: 8b9ad0f | NotFound++ | winxedst2.winxed: apply changes in stage 1 from the inline branch to stage 2 |
|||
| nxed: e1a409a | NotFound++ | winxedst2.winxed: add some binders and inline a few more things |
|||
| nxed: bedc804 | NotFound++ | winxedst1.winxed: fix merge master conflict |
|||
| nxed: 58b22fd | NotFound++ | winxedst2.winxed: diagnose inline with local functions not allowed |
|||
| nxed: a669e35 | NotFound++ | winxedst2.winxed: generic diagnostic for inline not allowed |
|||
| nxed: d431d64 | NotFound++ | winxedst2.winxed: diagnose yield inline |
|||
| nxed: c45da47 | NotFound++ | winxedst2.winxed: implement clone is some more operators |
|||
| nxed: 2c11441 | NotFound++ | examples/fly.winxed: use inline in example fly |
|||
| nxed: 64dbc7e | NotFound++ | winxedst2.winxed: implement clone in MultiStatement and pirops |
|||
| nxed: b79afdc | NotFound++ | winxedst2.winxed: implement clone in do, while, break and continue |
|||
| nxed: 300598a | NotFound++ | winxedst2.winxed: style fixes |
|||
| nxed: 38dcd26 | NotFound++ | winxedst (2 files): fix a problem with local functions sharing used subids with its outers |
|||
| nxed: 5e03a27 | NotFound++ | winxedst2.winxed: use some more binders |
|||
| nxed: 5bedc59 | NotFound++ | NEWS: update NEWS |
21:35 | ||
|
21:37
lucian joined
|
|||
| dalek | nxed: 9300774 | NotFound++ | pir/winxed_compiler.pir: update installable compiler |
21:39 | |
| rrot: a9e6183 | NotFound++ | ext/winxed/compiler.pir: update winxed snapshot to 9300774687 - fix FunctionId emit problem |
21:49 | ||
|
22:11
jsut joined
22:50
lucian joined
23:19
contingencyplan joined
|
|||
| dalek | kudo/nom: a227e73 | jnthn++ | src/core/Str.pm: substr was rather wasteful before, and didn't give the optimizer half as much to go on as it could; this improves things a bit. Found while profiling Simple::Redis by Woodi++. |
23:19 | |
| dukeleto | ~~ | 23:33 | |
| dalek | kudo/nom: da4426a | jnthn++ | src/core/ListIter.pm: Some small optimizations to ListIter. Knocks about 10% more off Simple::Redis benchmark, which has iteration on hot path. |
23:46 | |
|
23:50
snearch joined
|
|||