|
Post closed tickets in your report. | Note: This channel is for our weekly status meetings (Tuesdays at 19:30 UTC); you probably want #parrot instead. | irclog: irclog.perlgeek.de/ Set by moderator on 3 June 2012. |
|||
|
04:01
benabik joined
04:03
benabik joined
12:49
kjs joined
|
|||
| kjs | REPORT FOR KJS: | 12:54 | |
| * much work on M1's code generator to generate M0. | |||
| * arrays and structs now work (but only x.y format, not x.y.z yet). | |||
| * function calls and returns /almost/ work. Need improvement of M0 interp. | |||
| NEXT STEPS: | |||
| * write tests | |||
| * annotate source code with splint annotations. | |||
| * implement enums | 12:55 | ||
| * fix x.y.z and multi-dimensional arrays. | |||
| * document the design of M1. | |||
| WISHLIST: | |||
| * finish spec of functions and parameters. | |||
| * can we have an "address-of" instruction (C's &x operator)? (for instance to take the address of a register?) If not, then pointers are pretty much impossible to implement, and structs/pmcs will be handled as references instead (java style). Likewise, a contents-of operator (C's *x). | |||
| * someone working on M0's JIT to make this system fast. | |||
|
13:02
bluescreen joined
13:50
jashwanth joined
15:22
kjs joined
16:19
jashwanth joined
16:43
lucian_ joined
17:37
nine joined
18:24
lucian_ joined
18:30
rurban joined
19:13
NotFound joined
19:16
moritz joined
19:21
whiteknight joined
|
|||
| whiteknight | WHAT I DID: | 19:24 | |
| * Continue work on the whiteknight/io_cleanup1 branch. Hit a few big milestones (building, able to run tests). | |||
| * Working with GSOC students | |||
| * Fixing bugs | |||
| WHAT I WILL DO: | |||
| * Keep working on the io_cleanup1 branch to fix tests and finish implementations | |||
| EOR | |||
| moritz | WHAT I DID: | ||
| * worked on wrapping up my Rakudo+Perl 6 exceptions grant | |||
| * discussed parrot stuff with brrt++ and others | |||
|
19:24
tadzik joined
|
|||
| moritz | EOR | 19:24 | |
| rurban | DONE: | 19:25 | |
| * happy to be back, now @cPanel/Houston in a CST/GMT-5 timezone, | |||
| getting paid by cPanel to work ~100% for the perl foundation (perl5 mostly though) | |||
| * clang asan tests | |||
| * installation tests | |||
| * read through whiteknights IO cleanup | |||
| TODO: | |||
| * come up with a short testcase for a perl6 splice [],1 asan heap-overflow | |||
| * getting accustomed to the new git workflow | |||
| * check whiteknights IO cleanup | |||
| WISHLIST: | |||
| * reenable native pbc tests. without tests arch-compat will bitrot, | |||
| as it already did for 32->64bit 0.4-1.0 until I revived it. | |||
| .EOR | |||
| benabik | DID: | ||
| * Added a method to Key to get the type | |||
| * Set up a setup.winxed and test directory for PACT | |||
| WILL DO: | |||
| * Blog | |||
| BLOCKERS: | 19:26 | ||
| * The "Honey Do" list | |||
| EOR | |||
| (Also: old blocker: getting lost in the Key code. It's complex where it should be simple and vice versa.) | |||
| NotFound | Quick report: new winxed syntax: prefix ':' for declaring var and for parameters in multis. | ||
| EOR | 19:27 | ||
| q1q | |||
| moritz just has a general question: how are our GSoC students doing? | |||
| whiteknight | moritz: 50/50 | 19:28 | |
| benabik | Doing alright. Off to a slightly slow start, but not getting behind I think. I keep having to do piles of small things around the house to prep for the kiddo. | 19:29 | |
| cotto | howdy | 19:31 | |
| NotFound | Hola | 19:32 | |
| nine | WHAT I DID: | ||
| * hopefully fixed the timer problems on windows | |||
| WHAT I WILL DO: | |||
| * finish my Bachelor's thesis | |||
| * find out what's necessary to get threads merged | 19:33 | ||
| EOR | |||
| moritz | it seems #ps time has arrived | 19:34 | |
| cotto | so it does | ||
| benabik | #ps? What? Who? Why? | 19:35 | |
| nine | So, what does keep us from merging threads at this point? | 19:37 | |
| cotto | what's blocking it on your end? | ||
| anything? | |||
| whiteknight | nine: I suspect just wide testing | 19:38 | |
| if we can build Rakudo on that branch on windows and linux, I think that's the big hurdle | |||
| moritz | is branch mostly synchronized with master? | ||
| regarding non-threads changes, that is | |||
| nine | no blockers on my end. I'll disable threads_io.t on Windows. Then everything should pass there as well. | 19:39 | |
| whiteknight | okay, great. If we can get a report of somebody building and testing Rakudo on windows with that branch, I think we start moving forward for merge | ||
| nine | moritz: last merge seems to have been on May 13th | ||
| whiteknight | We should talk to alvis, see if we want to do that before the next release or after | ||
| moritz | nine: thanks, that should be recent enough | 19:40 | |
| NotFound | My question has the same issue: Someone objects to merge the notfound/pbc_to_exe_winxed branch? | ||
| Blocker: testing with msvc | |||
| moritz | NotFound: does it cause any rakudo regressions? | ||
| whiteknight | I'll fire up my vm tonight and test as many things as I can | ||
| NotFound | moritz: it generates byte-by-byte the same result as the hand coded pir version, so no regressions on anything. | 19:41 | |
| moritz | NotFound: then I'm +1 on merging | ||
| NotFound | (unless somenone builds parrot with a broken custom build script instead of make, of course, but I don't think somenone does that) | 19:42 | |
| cotto | +1 to merge after sufficient cross-platform testing | ||
| it's great to see thing getting merged. notfound++ nine++ | 19:44 | ||
|
19:44
kjs joined
|
|||
| NotFound | BTW: I think it's time to start dropping support for antique C, notably de declaration-after-statement thing. If somenone wants to use MSVC, they can build as C++. | 19:45 | |
| s/de/the | |||
| whiteknight | it's a conversation we do need to have | ||
| nine | Are there any drawbacks for people who compile as C++? | 19:48 | |
| NotFound | They usually get better optimization, so they can get scaried by the speed ;) | 19:50 | |
| benabik | At least for g++, it can be pickier about types and casting. | 19:51 | |
| NotFound | Seriously: I don't know if someone has tried, all reports I know about building as C++ are from g++. | 19:52 | |
| benabik: I can fix all that, if someone has problems. | |||
| I already do it, in fact. | 19:53 | ||
| cotto | +1 to saying that parrot support msvc using c++, barring a good reason to say otherwise | 19:57 | |
| *supports | |||
| Let's test it for a bit and see where that takes us. If we don't run into any blockers, we'll make it official. | 19:59 | ||
| any other questions or thoughts? | 20:00 | ||
| whiteknight | nope | ||
| NotFound | We just need to drop the options that set old C compatibility options from configure. | 20:01 | |
| cotto | That's a branch I'll be glad to see. | 20:02 | |
| NotFound | And without effort, people will start inserting incompatibilities inconsciously ;) | ||
| cotto | let's call it a wrap | 20:07 | |
|
20:07
moritz left
20:32
NotFound left
21:28
bluescreen joined
22:26
kjs left
22:53
whiteknight joined
23:06
whiteknight joined
23:23
lucian_ joined
|
|||