weekly meeting at 18:30 UTC, Tuesday | logged at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 30 September 2008.
01:38 Whiteknight joined 07:57 particle joined 10:12 particle joined 11:07 rdice joined 11:26 rdice_ joined 12:46 Wknight8111 joined 14:02 PacoLinux joined 14:30 leo joined 14:59 particle joined 16:06 pmichaud joined 16:20 jhorwitz joined 16:40 DietCoke joined 16:43 moritz joined 17:42 kj joined 18:13 cotto joined 18:19 rdice_ joined 18:22 Wknight8111_ joined 18:28 chromatic joined, barney joined
Coke ~~ 18:30
jhorwitz right back at ya
chromatic Good $LOCALTIME all.
kj hiya
moritz hi
Wknight8111 hello
cotto hi
barney hallo 18:31
chromatic Allison is in an airport, so I'll paste her report.
pmichaud hello.
chromatic Resolved a number of post-merge MMD problems. Have a few left to go.
atching up on mailing list posts (while I'm between development branches).
particle will be in late, on phone
chromatic Added a first draft of a guide to creating, maintaining, and merging branches. Review/patches welcome.
barney?
barney Fiddled with class attributes in Pipp. 18:32
.eor
chromatic I fixed a handful of bugs and cleaned up some type conversion warnings in C.
I'm going to review our use of CONST_STRING to make sure we NEVER EVER RETURN THEM TO CALLING CODE because that creates weird string bugs. 18:33
EOR
Coke?
18:33 NotFound joined
Coke - pulled partcl out of the repository, partially to make sure parrot 18:33
can deal with other languages doing this closer to 1.0, and partially 18:34
to give myself a stable work environment.
- managed to get mdiep to commit something! =-)
- Repository/wiki/links to everything at code.google.com/p/partcl
- partcl now passing 3158 tests running the native test suite.
[exit]
chromatic cotto?
cotto * have all known bugs ironed out of PhpArray code
* code still needs some cleanup (but less than last week)
* next step is to get to know the PCT code and start implementing functions in src/common
* queue 2 questions
eor
chromatic japhb?
japhb chromatic: as usual, nothing to report 18:35
chromatic jhorwitz?
jhorwitz at ppw last weekend. talks went well.
not much coding, as the "use" rakduo segfault continues to block and baffle me. i really need to open a ticket to track it.
EOR
chromatic kj? 18:36
kj == work on PIRC
* rewrote the heredoc preprocessor; it now correctly parses multiple heredocs in a sub call.
* Seems to work fine now, on cygwin, linux and windows.
* cleaned up the macro preprocessor. this links to libparrot as well now.
* const-ed all pirc sources as much as I could.
* added a basic Makefile, which was fixed by chromatic++ so it works on Linux (but needs fixing on windows)
* fixed about all compiler warnings, leaving only some warnings in generated sources (lexer, parser)
* found the way to remove the warnings about 'inconsistent dll linkage', which is also usable for IMCC.
* thinking about how to combine the 3 processors, so using it is easier.
* need to check out how the headerizer tool works, so it can generate the headers.
.eor
cotto unqueue one question 18:37
chromatic moritz?
moritz nearly no tuits, nothing to report.
chromatic NotFound? 18:38
NotFound A few work on some tickets.
EOR
chromatic particle? 18:39
particle ~5m please
chromatic pmichaud?
pmichaud ** Rakudo spectest_regression: 204 files, 4380 passing 18:40
== Miscellaneous
: presented Rakudo stuff at Pittsburgh Perl Workshop
: cleaned up older compilers still using n_* opcodes
== PCT stuff
: worked with Tene++ to finish hllmagic branch and merge back into trunk
: things that were 'Foo::Bar' in PGE and PCT are now ['Foo';'Bar']
== Rakudo stuff
: removed some tests from spectest regression until we can fix in Parrot
: worked on improving package and namespace handling in Rakudo
: added 'parse_name' method to Perl6::Compiler
: fixed use of internal $?NS variable
: eliminated incorrect definition of '::' as a sigil
: working towards enabling interpolated namespaces (for mod_parrot)
: worked with particle++ on 'is export'
: added a couple of new features to progress-graph.pl
EOR
jhorwitz pmichaud++ :)
chromatic tewk? 18:41
Tene? 18:42
Tene A little work on cardinal. 18:43
namespace updates with pmichaud
Started trying to work on a branch to get languages into their own .HLLs, but blocking on some feedback/decisions from pmichaud.
KTHXBAI
chromatic particle?
particle ~ attended pghpw, hacked with pmichaud, jhorwitz and others
~ met folks from buglabs.net
~ got a hacker to start work on porting parrot to openembedded linux (which the bug runs)
~ got 'is export' trait to start to work in rakudo 18:44
~ some funny bugs are preventing me from continuing on that
~ i'll bend patrick's ear about them
~ hopefully we can get past them quickly
~ discussed branch merge philosophy with will and allison
~ working on the parrot developer summit logistics
1 queuestion :)
.end
chromatic Wknight8111? 18:45
Wknight8111 * Worked on calling_conventions branch some.
* Been busy in real life (always seem to be busy).
EOR
chromatic Let's move to questions. cotto had one and particle has one. 18:46
cotot?
cotto?
cotto Could someone review the fix in rt.perl.org/rt3/Ticket/Display.html?id=59810 and let me know if it's the right approach?
Summary: Because STRINGs don't cache the seed used to compute a hash, it's possible to break stuff by computing (and therefore caching) the hashval of a STRING using a seed other than the one used in src/hash.c.
The patch in that rt adds code to cache and check the seed so API users can use an arbitrary seed without breaking anything.
Coke has one now, too.
chromatic cotto, it seems like it works, but I'm not sure it's the cleanest solution. I'm not sure what the cleanest solution is, however. 18:47
18:48 Wknight8111_ joined
Wknight8111_ did my report get through? 18:48
cotto should I just give it time to see if we can find something better?
Wknight8111_, yes
particle Wknight8111_: only the part up to EOR
Wknight8111_ ok 18:49
chromatic cotto, how much breaks without it?
cotto just my phparray code, and there's an easy workaround 18:50
I'm more concerned about future mysterious breakage.
chromatic Me too. 18:51
The idea of having the cached hash value flip flop back and forth seems a little silly to me, and the idea of expanding our string struct to include even more data... but for longer strings, it may be worthwhile. 18:52
cotto What about not allowing the seed to be changed (apart from recompiling)? 18:53
chromatic Long-term, I'm not sure that's useful either.
There are some hash-poisoning effects that can make really long bucket chains that create easy DoS attacks. Changing the seed dynamically is a nice way to avoid that. 18:54
particle perl 5.8.4 is the one that finally got that right iirc
chromatic Let's leave it as a TODO for now, if you have a workaround and it isn't blocking you. 18:55
Will that work?
cotto that's fine for now
chromatic particle, you had a question.
particle release time is coming up. does anybody have issues that need priority before release? i'm sure the answer is yes. 18:56
if so, please add them as dependencies to the release tracking ticket in rt#58946
and, please, let's look at fixing them, too :)
chromatic Fixing TGE! 18:57
Coke, you had a question. 18:58
Coke (and Tcl/Glob, which was my question, and I just added the tcl (59870) as a dep to 58946).
added the -ticket-
particle coke+=1
chromatic cokf 18:59
Coke It would be nice ot have this release be one I can target with the new out-of-the-repo partcl.
.
chromatic Are there any other questions?
That looks like a no. 19:01
Gentle reminder: if everyone closed one RT a day until the next release, we'd be under 600 open bugs.
Wknight8111 feels inspired! 19:02
chromatic GO HIT RESOLVED READY... BREAK! 19:03
19:12 chromatic left 19:17 Coke left 19:19 jhorwitz left 19:22 pmichaud left, davidfetter joined 19:30 moritz left 20:15 particle1 joined 20:30 PacoLinux left, particle joined 21:06 particle1 joined 21:13 particle joined 22:03 Whiteknight joined

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!