Parrot 3.5.0 "Menelaus" released | parrot.org | Log: irclog.perlgeek.de/parrot/restoday
Set by moderator on 25 June 2011.
dalek kudo/nom: 1357b60 | jonathan++ | src/ (5 files):
Map Parrot RPA and Hash into Perl 6 types after method calls, so we can call a bunch more stuff from NQP land. Makes the various intropsection methods that didn't work previously now do so. A few things relied on this not happening in list guts; patched them.
 00:27
kudo/nom: ab2b93b | jonathan++ | / (2 files):
Implement .+ and .*.
kudo/nom: 8c0cbc0 | jonathan++ | / (2 files):
Make literal values in signatures work again.
00:29 Coke_ joined
Coke_ is finally back online from home. 00:29
00:29 coke left 00:30 Coke_ is now known as Coke
whiteknight Coke++ 00:30
dalek website: soh_cah_toa++ | It's a Miracle!
website: www.parrot.org/content/its-miracle
whiteknight "All the other pieces of the puzzle should now slip right into place nice and easy. Good times lie ahead." 00:32
obviously soh_cah_toa is not an experienced developer!
soh_cah_toa ha!
whiteknight Don't worry though, the cynicism comes with experience 00:33
dalek rrot-gmp: 7104323 | bubaflub++ | / (6 files):
update blacklist, fix script, regenerate bindings
 00:40
rrot-gmp: 0265cb0 | bubaflub++ | gmph2ncidef.pl:
fix gmph2ncidef - use non-capturing group
rrot-gmp: 1592c44 | bubaflub++ | / (4 files):
update bindings to include gmp_rand* funcitons
bubaflub whiteknight: can you help me with a C question? 00:48
whiteknight yessir 00:49
bubaflub in GMP land there is a struct called gmp_randstate that holds all kinds of information related to random number generation. first, the struct looks like this:
whiteknight I'm always happy to help. You don't need to ask. Just say "damnit Whiteknight. get over here and help me!"
bubaflub typedef struct
{
mpz_t _mp_seed;\t /* _mp_d member points to state of the generator. */
gmp_randalg_t _mp_alg; /* Currently unused. */
union {
void *_mp_lc; /* Pointer to function pointers structure. */
} _mp_algdata;
} __gmp_randstate_struct;
whiteknight oh, pretty
bubaflub so my first question is why the union around a single pointer? 00:50
whiteknight that's a very good question.
bubaflub is it just a pointer then? i'll eventually need to represent this struct with a StructView 00:51
whiteknight yes, that would collapse down to be just a pointer 00:52
I don't think it does anything else. The union shouldn't affect the alignment of it or anything
bubaflub ok, great. second question - the gmp_randalg_t is defined as an enum like this:
typedef enum
{
GMP_RAND_ALG_DEFAULT = 0,
GMP_RAND_ALG_LC = GMP_RAND_ALG_DEFAULT /* Linear congruential. */
} gmp_randalg_t;
whiteknight It might be something for backwards compatibility, like they used to have more things in there, but deleted them
bubaflub whiteknight: that would make sense.
whiteknight in C, enums are ints
bubaflub whiteknight: great. i'll never be accessing this information directly, just shuffling data in and out of a (hopefully) properly designed StructView 00:53
whiteknight right
bubaflub whiteknight: ok, last question - after both of those definitions there is a typedef like this:
typedef __gmp_randstate_struct gmp_randstate_t[1];
i've never seen something typdef'd to an indexed array
what does that do?
whiteknight ah, tricky. Don't see that much 00:54
gmp_randstate_t is a 2-element array of __gmp_randstate_struct
in the win32 API you see a lot of things like "typedef Foo *pFoo", where pFoo is typedef'd to be a pointer to Foo
bubaflub with both elements being of the type gmp_randstate_t?
whiteknight this is the same thing. Think about this like an ordinary declaration 00:55
ie, get rid of the "typedef" and then read it
yes
bubaflub ok, that makes more sense. i've usually seen typedefs with simple types 00:56
whiteknight yeah, these kinds of uses are not very common
dalek rrot/soh-cah-toa/hbdb: d224aec | soh_cah_toa++ | / (3 files):
Started organizing Parrot::Test::HBDB into an OO-interface. It's not finished but I need these changes at YAPC::NA
whiteknight msg plobsing I'm working on a new packfile pmc in whiteknight/packfilewrapper branch. Do you think it's worthwhile to expose PackFile_append_pmc functionality to the user? 01:02
aloha OK. I'll deliver the message.
dalek rrot/whiteknight/packfilewrapper: 920d057 | Whiteknight++ | / (4 files):
misc fixes and cleanups. +docs. Add new Parrot_pf_serialized_size API and create method PackfileView. Use more packfile API funcs from the embedding API
 01:21
whiteknight msg cotto before I do any more work on whiteknight/packfilewrapper branch, can you take a look at it to make sure it's sane and I'm moving in an acceptable direction? Thanks! 01:23
aloha OK. I'll deliver the message.
cotto whiteknight, deal 01:25
whiteknight no rush, I've got a million other things to play with 01:28
bubaflub whiteknight: how does 'using namespace' in Winxed work? does it load all the junk from another namespace in? 01:34
whiteknight bubaflub: I don't know exactly. I think it's a compile-time thing 01:37
so the compiler will search in other namespaces without having to qualify each call
I think
bubaflub whiteknight: ok, i think that's what i want. i want to move a lot of my StructView stuff and string utility stuff out of GMP::Integer into GMP::Common cause i'll need it in every GMP module 01:38
ping NotFound 01:39
whiteknight he might be asleep now 01:41
aloha time?
aloha whiteknight: time is a lion, where i am a lamb.
whiteknight damnit
aloha clock?
aloha whiteknight: LAX: Sat, 18:41 PDT / CHI: Sat, 20:41 CDT / NYC: Sat, 21:41 EDT / UTC: Sun, 01:41 UTC / LON: Sun, 02:41 BST / BER: Sun, 03:41 CEST / TOK: Sun, 10:41 JST / SYD: Sun, 11:41 EST
whiteknight yeah, he's in Spain, so near UTC
sorear whiteknight: your vectorized tests remind me of TestML 01:51
whiteknight never used TestML
sorear neither have I
but I want to migrate roast to a similar system 01:52
whiteknight ok
I've got a million ideas for things to add to the test library. Adding things is more slow than imagining them 01:53
01:54 contingencyplan left
dalek rrot: d3edfd8 | Whiteknight++ | include/parrot/hash.h:
Merge branch 'whiteknight/remove_hash_macros'
 02:05
rrot: 9b37515 | Whiteknight++ | src/call/args.c:
Fix two instances of old-style hash function names
ttbot Parrot d3edfd85 i386-linux-thread-multi make error tt.taptinder.org/cmdinfo/26449 02:08
NotFound bubaflub: pong 02:09
bubaflub NotFound: can you describe what 'using namespace' does? 02:12
NotFound bubaflub: it makes available the content of the namespace from the current scope. 02:13
02:13 daniel-s left
bubaflub NotFound: so if i functions in GMP::Common and i do a 'using namespace' in GMP::Integer i don't have to qualify my functions, right? 02:14
02:14 gbacon_ joined
bubaflub i.e. i could call func_name instad of GMP.Common.func_name 02:14
NotFound bubaflub: yes, but keep in mind that is a compile time feature. 02:15
bubaflub NotFound: what exactly do you mean? 02:17
02:17 gbacon_ left
NotFound bubaflub: it makes available only symbols present in the current source file, doesn't discover things compiled apart. 02:19
bubaflub NotFound: ok, so one namespace per file? or one class per file? 02:20
NotFound You have example usages in t/advanced/10scope.t 02:21
bubaflub ok, i'll take a look. thanks. 02:22
NotFound On the contrary, it works only within namespaces contained in one file.
Some way of including headers and declaring extern files is planned, but not ready yet. 02:23
bubaflub NotFound: hmmm. ok, what i'm trying to do is pull out all of my StructView and string functions into GMP::Common. how should i declare that stuff, and how should i include it in GMP::Integer (and eventually other GMP::* classes)? 02:24
right now i have namespace GMP class common in one file and namespace GMP class Integer in another file 02:25
NotFound You can use 'using extern', but that is a different beast: it imports into the current namespace at runtime.
bubaflub what would be cleanest? 02:27
NotFound bubaflub: there is not a really clean way right now. 'using' at function scope is the simpler way, but verbose. 02:28
bubaflub NotFound: no big deal - most of the code is generated so verbose ain't bad 02:29
thanks.
02:38 rurban_ joined
whiteknight LexInfo appears to be a very thin wrapper around Hash 02:38
we can probably deprecate+remove it without hurting anybody
02:39 daniel-s joined 02:41 rurban left, rurban_ is now known as rurban
whiteknight if we add a "keys()" method to return an array of keys to Hash, we can kill LexInfo completely and just use hash for that purpose 02:42
NotFound whiteknight: I think it's needed, even if only to be able to HLL map it.
whiteknight ah, that is a good point 02:43
NotFound BTW, I think both keys() and values() methods on Hash will be useful. 02:44
whiteknight yes 02:46
I'm still not certain I understand the need for LexInfo at all. I don't see why we need that functionality
NotFound Better ask nqp and rakudo guys. 02:47
whiteknight Why not let LexPad be an ordinary hash with no restrictions? Make any named lexical you want without having to declare them at compile time
HLLs can add their own restrictions if they want them
02:47 pbaker1 joined
cotto goes 02:48
02:50 pbaker left 02:54 cotto left 02:59 whiteknight left
dalek rrot-gmp: a059520 | bubaflub++ | / (3 files):
two more random functions
 02:59
rrot-gmp: 180db54 | bubaflub++ | / (7 files):
create GMP::Common to hold common bits between all GMP classes

fix gmpdoc2winxed to generate the proper 'using' statements add GMP::Common and generated pir update to GMP::Integer template, winxed, and generated pir fix one test that was relying upon string utility functions now in GMP Common
bubaflub msg plobsing i've got a few questions about StructView stuff typed up here: gist.github.com/1047177 - any help would be most appreciated. 03:08
aloha OK. I'll deliver the message.
bubaflub msg dukeleto bloggity blog blog: www.parrot.org/content/parrot-gmp-e...d-nci-work 03:34
aloha OK. I'll deliver the message.
03:37 Coke left
dalek website: bubaflub++ | Parrot-GMP: Even more tests and NCI work 03:37
website: www.parrot.org/content/parrot-gmp-e...d-nci-work
03:37 Coke joined
sorear I am curious whether Rakudo will be able to use Parrot-GMP anytime in the next, say, 2 years 03:37
bubaflub sorear: i hope so. 03:38
sorear: and i'm not much in Rakudo land but i'm willing to help. 03:39
sorear history tells me that Rakudo can use something if and only if it is bundled with Parrot
Rakudo hasn't supported ICU on all platforms since it was debundled
sorear is pessimistic today. 03:40
bubaflub sorear: well, i don't know if it'd be useful bundling directly with Parrot though it's pretty light-weight given that Winxed is bundled - i can generate all the files from there. 03:41
03:54 cotto joined 04:00 theory joined
dalek kudo/nom: 1cb0fb2 | pmichaud++ | src/ (5 files):
Avoid need for nqp::find_method calls on RPAs.
 04:35
cotto soh_cah_toa, ping 04:51
soh_cah_toa cotto: pong
cotto soh_cah_toa, does t/tools/hbdb/bad_cmds.t run on your machine?
soh_cah_toa cotto: w/ the most recent changes, it probably shouldn't 04:52
cotto soh_cah_toa, ok. just checking if that's expected
because it doesn't
soh_cah_toa it's b/c i'm reorganizing Parrot::Test::HBDB and moved some things from there into the package
cotto also, "help" doesn't live up to its name. 04:53
wfm
how much of that code is cargo culted from other places?
soh_cah_toa the tests or source?
cotto the tests 04:54
soh_cah_toa hmm...none. i think.
i know i used pbc_disassemble as a guide on one of them though
cotto so you understand all the bits in Parrot::Test::HBDB? 04:55
soh_cah_toa yeah
well except for one
which i want to ask
cotto import?
soh_cah_toa the Test::Builder. let me find it
ha! yes
the export_to_level()
cotto I'll have to find that out for myself first. 04:56
soh_cah_toa that part i copied/pasted from some other parrot module
cotto that's the one that stood out. 04:57
blessing a hashref == ouch, but we don't require moose 04:58
soh_cah_toa huh? that's how you're supposed to do it
cotto That's the primitive. It's better to use a proper object system if you have a choice. 04:59
i.imgur.com/AKUiT.jpg
soh_cah_toa ha! 05:00
cotto export_to_level is part of Exporter, which I don't understand especially well.
soh_cah_toa apparently, you use it if you use Exporter but also want your own import method 05:01
the perldoc has 1 as the first argument which makes sense but Parrot::Test uses 2 so that's what i kept. whether 2 or 1, it didn't seem to make a different oddly enough 05:02
soh_cah_toa will brb 05:05
cotto soh_cah_toa, I'll buy you a beer if that image doesn't appear somewhere at yapc. 05:07
soh_cah_toa are blessed hashrefs really that frowned upon? i mean, in every single perl 5 book i have that mentions something about oo perl, bleshed hashrefs are there 05:20
sorear blessed hashrefs are the de facto standard way to do objects in Perl 05:21
soh_cah_toa that's what i thought
cotto boarding 05:28
see you at yapc 05:29
soh_cah_toa yeah, i'm leaving in an hour
see ya
05:35 cotto left
dalek kudo/nom: 0b11f42 | pmichaud++ | src/core/ (2 files):
Temporarily revert 348b506 -- something about $*OUT.say doesn't work more than once in the interactive REPL.
 05:51
06:00 soh_cah_toa left
Coke rakudo doesn't support IMCC if it's not bundled with parrot likely because of the way parrot configs for IMCC to start with. 06:00
sorear Coke: do you mean ICU? 06:03
06:04 particle1 joined
sorear finds the notion of Rakudo not supporting IMCC hilarious 06:04
06:05 particle left
Coke whoops. yes. 06:05
waaay past my bedtime.
yay. now that I can ssh back home when needed, I can restart my taptinder client. 06:06
sorear unbundled IMCC seems to be a common fantasy around here 06:11
Coke eh. it started unbundled, I'm sure it'll end up that way. whiteknight seems to want it so. 06:18
06:29 rohit_nsit08 joined
dalek rrot-gmp: d8f3cab | bubaflub++ | gmpdoc2winxed.pl:
take regex prefix as a command line option and fix internal name
 06:29
rrot-gmp: 451da97 | bubaflub++ | src/GMP/Common. (2 files):
first shot at getting gmp_randstate_t struct to work
rrot-gmp: 0dc2da0 | bubaflub++ | src/GMP/Random. (3 files):
GMP Random initialization functions template, Winxed with docs, and generated PIR
rrot-gmp: 198f579 | bubaflub++ | t/random/init.t:
first test for random functions
rrot-gmp: cf2303a | bubaflub++ | t/harness:
update harness for random directory test
kudo/nom: e333382 | moritz++ | / (3 files):
implement .chop, update LHF
 06:41
06:41 dodathome joined 06:43 theory left 06:45 theory joined, Eclesia joined
dalek rrot-gmp: 1ca2e46 | bubaflub++ | t/random/init.t:
tests for randinit functions

this covers gmp_randinit_default, gmp_randinit_mt, gmp_randinit_lc_2exp, and gmp_randinit_lc_2exp_size		 06:48
rrot-gmp: c440cb3 | bubaflub++ | t/random/set.t:
test for gmp_randinit_set
rrot-gmp: d117b64 | bubaflub++ | t/random/clear.t:
test for gmp_randclear
rrot-gmp: ced52e0 | bubaflub++ | t/random/seed.t:
tests for randseed functions

this covers gmp_randseed and gmp_randseed_ui
kudo/nom: aa21eb7 | pmichaud++ | / (3 files):
Get a basic try { ... } working -- catches exceptions and puts them into $!.
 06:52
kudo/nom: bdbd8d7 | pmichaud++ | src/core/Str.pm:
Add a XXX note to Str.chop since it ends up sending negative lengths to .substr
 06:54
kudo/nom: 3fe16cb | moritz++ | src/core/Str.pm:
fix chop, pmichaud++
 07:10
07:10 theory left 07:21 zby_home_ left 07:25 daniel-s left, daniel-s_ joined 07:30 daniel-s__ joined 07:32 daniel-s_ left 07:36 _daniel-s__ joined 07:37 daniel-s__ left 07:41 daniel-s joined 07:43 _daniel-s__ left 07:46 daniel-s_ joined 07:48 daniel-s left 08:14 jsut_ joined 08:19 jsut left
Eclesia hello 08:23
is there a winxed command to load/unload a *.pbc ? 08:24
not talking about : $load 'something.pbc';
09:37 contingencyplan joined 09:50 contingencyplan left 09:55 whiteknight joined 09:57 mj41 joined
whiteknight good morning, #parrot 09:59
Eclesia hi whiteknight
whiteknight: I have a question for you :) 10:00
whiteknight Hello Eclesia, how are you today?
Oh good! I'm filled with answers!
Eclesia I defined a namespace : namespace "{eria.model}parser" {
but I can reuse it from another class 10:01
whiteknight what do you mean?
Eclesia tryed : using {eria.model}parser.Compiler;
or : var compiler = new eria.model}parser.Compiler(); 10:02
+{
but looks like winxed do not like { } in namespace names
and I can't find a working escape caracter
whiteknight There is an alternate syntax you can try: 10:03
10:03 JimmyZ joined
whiteknight new ["{eria.model}parser", "Compiler"] 10:03
10:06 cjh left
Eclesia whiteknight: thanks it passes compilation now 10:06
whiteknight awesome
Eclesia whiteknight: but complains about something else now :x 10:07
Class '[ '{eria.model}compiler' ; 'Compiler' ]' not found
is the -L option recursive ?
whiteknight I don't think so
Eclesia whiteknight: is that correct ? 10:23
nopaste "Eclesia" at 192.168.1.3 pasted "// Eria base library $load '{e" (13 lines) at nopaste.snit.ch/55132
whiteknight That looks right 10:25
Eclesia hm ...
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#16385), fulltest) at 3_5_0-39-g9b37515 10:26
Ubuntu 11.04 i386 (g++)
dukeleto waves from the Newark airport 10:31
Eclesia whiteknight: you are only linux ? 10:32
on*
whiteknight Eclesia: on the weekends, yes
dukeleto: You're in Newark? 10:33
dukeleto whiteknight: yeah, heading to YAPC::NA
whiteknight: me and cotto are on the same flight, hilariously
whiteknight dukeleto: Oh, so you're just on layover? 10:34
dukeleto whiteknight: yeps
whiteknight: i wouldn't be here for any other reason :)
dukeleto left these parts long ago for a good reason
whiteknight I would have driven down to meet you guys if I had known 10:35
start paying up on all those beers I owe you two
Get you more drunk than the folks in first class
10:35 mj41 left
dukeleto whiteknight: :) . We do indeed need to meet up for delicious beverages sometime soon 10:36
Eclesia whiteknight: can I ask you to compile my project on your pc ? I can't find what's wrong ...
whiteknight Eclesia: Sure thing. Where can I get a copy?
10:36 cjh joined
Eclesia whiteknight: ftp://ftp-developpez.com/jsorel/temp/eria-pvm.zip 10:36
whiteknight I'm actually excited to see it
Eclesia lol there is nothing much to see ^^
only one week (and only evenings) on it 10:37
dukeleto clones winxed.git for the first time
Eclesia whiteknight: run the compile.sh in the root folder
whiteknight I'm building parrot right now, I'll look at it in a little bit
10:37 ligne joined
Eclesia whiteknight: np 10:38
10:38 rurban_ joined 10:41 rurban left, rurban_ is now known as rurban
dukeleto finds a buglet in Winxed: "parrot setup.pir test" assumes/run an installed winxed 10:43
msg NotFound github.com/NotFound/winxed/issues/1 10:45
aloha OK. I'll deliver the message.
Eclesia how can I print all available namespaces and classes ? 10:46
dalek rrot: 47c310c | dukeleto++ | NEWS:
Add a note to NEWS about winxed.git
 10:48
dukeleto Eclesia: hmmm. Good question.
dukeleto clones and compiles Rosella for the first time 10:51
whiteknight: i am seeing a test failure in t/memoize/Memoize.t in Rosella. Known issue? 10:58
whiteknight nope
I'll look at it in a minute
dukeleto whiteknight: github.com/Whiteknight/Rosella/issues/15 11:02
whiteknight dukeleto++
Eclesia: winxed bug
in Setup.wx, do this: var compiler = new ['"{eria.model}compiler"', "Compiler"]();
Eclesia damn 11:03
whiteknight the namespace name literal has the quotes in it, for some reason
do that and it works
dukeleto whiteknight: just added the sha1's of parrot and winxed that caused that, to the issue
Eclesia whiteknight: ok thanks, I'll do with it until it's fixed 11:04
dukeleto whiteknight: sadly, github markdown tried to be smart and autolink to sha1's in rosella.git
whiteknight: i will update to the latest parrot and see if I can reproduce
11:06 jsut joined, jsut_ left
whiteknight dukeleto: how old was your parrot? 11:08
dukeleto whiteknight: 3.4.0-ish 11:12
11:12 jsut_ joined
whiteknight hm, I don't think that test relies on any modernish features 11:14
NotFound whiteknight: What's the bug?
whiteknight NotFound: namespace "Foo" { }
then you have to call it ['"Foo"'] 11:15
namespace "Foo" { function bar() { say(__FUNCTION__); } }
that causes a syntax error, because __FUNCTION__ contains ""Foo"".bar
11:16 redicaps joined
whiteknight that causes a PIR error 11:16
11:17 jsut left
NotFound Uh.... funny. The bug is to not diagnose the syntax error, namespaces names that are not identifiers are not supported yet. 11:17
whiteknight hah, okay 11:18
NotFound: how hard is it to support names that are not identifiers?
NotFound Not sure, doing it badly can complicate the parser at several points. 11:20
whiteknight ok
Eclesia isn't it just removing extra "" written in the pir file ? 11:21
11:24 cotto joined, cotto left 11:25 cotto joined
NotFound Eclesia: Do you have any complelling reason to use that kind of names? 11:25
Eclesia that's a module name. it's a comon used writing for a Qname in extended form {namespace}local 11:26
If you know maven, it's similar to groupId + artifactId 11:28
11:29 cottoo joined
Eclesia NotFound: what is the pattern for a valid identifier ? maybe I could find something approching 11:30
cottoo whiteknight, you around? 11:31
whiteknight yessir
NotFound Eclesia: alphanumeric, _ and $, but $ can't be the first character,
cottoo I took a look at your branch on the way over. It's looking pretty good.
I do have some questions though 11:32
NotFound Numeric also can't be the first.
11:32 redicaps1 joined
cottoo Is there a reason to treat :init and :main as special in the PackfileView PMC? istr a plan to make those more generic. Having trigger_init, trigger_load and main_sub feels clunky. 11:32
whiteknight cottoo: my only motivation was to provide access to current functionality. I could try to make it more forward-thinking 11:34
11:34 cotto left
whiteknight main_sub is necessary, but the other two could be updated 11:34
something like "trigger(<name>)", where "load" and "init" are the only two available options right now 11:35
cottoo that's what I was thinking
dukeleto whiteknight: Memoize.t still fails for me on the latest winxed.git built with the latest parrot.git
whiteknight dukeleto: weird.
dukeleto whiteknight: let me know if I can give you more details
whiteknight dukeleto: I'll need to try to get the value that's being returned 11:36
Eclesia NotFound: the issue is a parrot limitation of it's winxed parser ?
or*
NotFound Eclesia: winxed parser and code generation.
whiteknight dukeleto: are you on a 32-bit machine?
11:37 redicaps left
whiteknight I wonder if the answer there is overflowing 11:37
cottoo whiteknight, What's pfview.all_subs' intended use case? 11:38
11:38 cottoo is now known as cotto
dukeleto whiteknight: i am printing the number and it is the same. Seems like it is using object equality instead of integer equality 11:38
whiteknight cotto: a replacement for the same functionality in Eval
Eval.elements provides the number of Subs, and Eval.get_pmc_keyed_int returns the Sub constant at that index 11:39
cotto ok
whiteknight so, instead of doing that internally to PackfileView, since I want to support all constants, I return an array of Subs
dukeleto whiteknight: i can haz commit bit to Rosella?
whiteknight dukeleto: yessir
dukeleto: Definitely overflows INT_MAX on 32-bit machines 11:41
I'm going to assume you have GMP, so it's probably autopromoting to BigInt
dukeleto whiteknight: yep, i have GMP in my parrot 11:42
cotto whiteknight, my big concern is that overall, the naming seems a bit rushed. +meh to "Parrot_pf_write_pbc_file" and pfman.write_to_file
otherwise, the branch is awesome. 11:43
whiteknight++
dukeleto whiteknight: an easy test for memoization is to memoize a function that returns the current time
11:51 ligne left 11:52 Eclesia left
whiteknight cotto: what names would you prefer? 11:54
dukeleto whiteknight: if you flip my commit bit I will hack on some Rosella tests :) 11:55
whiteknight I'm working on it. My mouse driver is wonky after an update last week, and it's hard to navigate github without it
added 11:56
dalek sella: 3b416df | dukeleto++ | t/memoize/Memoize.t:
Add a useful diagnostic to a Memoize test that fails on my machine
 11:57
whiteknight :)
dukeleto whiteknight: rosella/test.winxed seems to force object equality, even if given strings/numbers 12:00
whiteknight: $!assert.equal() does, that is
whiteknight I figured that was the most general way, at least until winxed supports MMD 12:01
dukeleto whiteknight: seems reasonable 12:02
whiteknight: so the memoize tests pass for you? I wonder what is causing the difference
whiteknight I'm on a 64-bit machine without GMP installed
cotto whiteknight, s/Parrot_pf_serialize_to_string/Parrot_pf_serialize/ s/Parrot_pf_deserialize_from_string/Parrot_pf_deserialize/ 12:03
whiteknight so those values are both Integer
cotto: noted
cotto the "from_file" and "to_file" in packfilemanager seem redundant, but that might be the red-eye talking 12:04
whiteknight dukeleto: what is the result ofsay(typeof(12586269025)) on your system?
or say(12586269025 == 12586269025)? 12:05
dukeleto whiteknight: typeof says it is an Integer 12:08
whiteknight: i had to use pir::typeof(), is that correct? 12:09
whiteknight: how do I use dumper.pbc from NQP? I can load dumper.pbc, but pir::_dumper() doesn't seem to work
whiteknight I think it's just "_dumper()" 12:10
dukeleto whiteknight: indeed it is. say(12586269025 == 12586269025) prints 1 12:11
whiteknight: yes, overflow is the problem 12:13
whiteknight: dumping the integer gives "VAR1" => -298632863
whiteknight: so NQP is overflowing the int, which causes the test to fail 12:15
cotto adios 12:16
dukeleto gets ready to board the plane to YAPC with cotto 12:17
12:22 cotto left
dalek sella: 7067661 | dukeleto++ | t/memoize/Memoize.t:
Fix #15 by passing a string to assert.equal, instead of a large integer, which overflows on 32bit systems
 12:30
12:36 ambs joined 12:43 lucian joined
whiteknight dukeleto++ 12:50
13:07 bacek_at_work left, aloha left 13:09 aloha joined, bacek_at_work joined 13:11 szabgab left
dalek kudo/nom: 1d7e3cd | jonathan++ | Test.pm:
Sprinkle protos through Test.pm to get us a bit further with compiling it.
 13:24
kudo/nom: 7555814 | jonathan++ | src/Perl6/Metamodel/BOOTSTRAP.pm:
Missing decontainerize.
kudo/nom: 0018881 | jonathan++ | src/Perl6/SymbolTable.pm:
Few tweaks so using Perl6::SymbolTable from Perl 6 code works out better.
kudo/nom: 1c5e0dd | jonathan++ | src/Perl6/Actions.pm:
Fix issue where we set name too late, so the compile time created Code object's .name didn't work.
kudo/nom: 9cc88cc | jonathan++ | src/ (3 files):
First cut of 'is export' trait; just implemented for routines for now.
kudo/nom: ec37e1b | jonathan++ | / (3 files):
Add Callable role; make Code do it.
 13:37
pmichaud good morning, #perl6 13:50
oops, wrong chan. but good morning anyway :) 13:51
dalek kudo/nom: 4fbc478 | jonathan++ | src/Perl6/Actions.pm:
Fix modifier handling crash and eliminate another bit of Q:PIR as a bonus.
 13:53
kudo/nom: 810c3b9 | jonathan++ | Test.pm:
Comment out a few small bits of Test.pm that causes us problems at the moment. Now it compiles and seems to kinda work ('use Test; ok(1); done;' is fine).
13:57 ligne joined
dalek kudo/nom: 97c00e5 | pmichaud++ | src/core/Str.pm:
Avoid extra box/unbox of constant integer literal to pir::chopn.
 14:14
whiteknight good morning, pmichaud 14:30
pmichaud good morning, whiteknight 14:31
14:35 rohit_nsit08 left
bubaflub morning, #parrot 14:43
dalek kudo/nom: 76d1efc | pmichaud++ | src/ (2 files):
Move exception creation code for <try> into a

catch anything.		 14:45
tadzik good afternoon parrots 14:46
14:46 zby_home joined 14:54 fperrad joined 14:56 fperrad left 14:58 fperrad joined 15:02 redicaps1 left
whiteknight hello bubaflub, tadzik 15:05
how go the GSoC projects?
bubaflub whiteknight: good - i did a minor refactoring last night and added support for random number generator functions 15:06
whiteknight oh nice
bubaflub i don't think i have the StructView quite right, but i'll harass plobsing whenever he's around next
tadzik no work here, exams on my head 15:08
whiteknight tadzik: still exams? 15:11
tadzik aye 15:12
NotFound dalek is not looking at new winxed repo yet. It's automatic or someone must tell it to reconfig? 15:13
whiteknight NotFound: you need to add a post-commit URL
tadzik your favourite signal theory is tomorrow
NotFound whiteknight: add where? 15:14
whiteknight NotFound: "Admin". There should be a tab for post-commit actions. Find the section for Urls
"Admin" > "Service Hooks" 15:15
Add this Post-Commit URL: host04.appflux.net:5000/dalek?t=magnet,parrot
lucian whiteknight: i'm just recently back from a (short) vacation 15:18
being lazy helps
whiteknight yes. Vacations can be very invigorating
lucian aloha: tell cotto i'll consider downscaling the puffins. but they're so cute! 15:19
NotFound Done
dalek nxed: 1a8a808 | NotFound++ | setup.json:
update repository uris
nxed: f7c324c | NotFound++ | winxed.json:
delete plumage metadata
nxed: 75ad7e2 | NotFound++ | t/advanced/03extern.t:
tests for 'extern function'
NotFound It works, thanks.
whiteknight: BTW, take a look at the new feature 'extern function' 15:23
whiteknight NotFound: that's exactly what I needed 15:24
awesome
dalek nxed: 133ce30 | NotFound++ | pir/winxed_compiler.pir:
update instalable compiler
 15:28
kudo/nom: 5ee2639 | jonathan++ | src/Perl6/Actions.pm:
Fix bug involving inline blocks not being executed.
 15:29
kudo/nom: fe6f8cf | jonathan++ | src/ (2 files):
Mu.isa and Mu.does.
15:29 daniel-s_ left
dalek kudo/nom: 7006946 | pmichaud++ | src/ (3 files):
Merge branch 'nom' of github.com:rakudo/rakudo into nom
 15:30
rrot: ffe78e1 | NotFound++ | ext/winxed/compiler.pir:
update winxed snapshot to get recent features, mainly 'extern function'
 15:33
15:34 soh_cah_toa joined 15:35 soh_cah_toa left
dalek kudo/nom: 0d75a77 | jonathan++ | src/Perl6/SymbolTable.pm:
CHECK and END are LIFO (pmichaud++).
 15:37
kudo/nom: fb649ae | pmichaud++ | src/core/Mu.pm:
Add prefix:<defined>.
 15:40
kudo/nom: 196d776 | pmichaud++ | src/Perl6/SymbolTable.pm:
Merge branch 'nom' of github.com:rakudo/rakudo into nom
15:44 contingencyplan joined
dalek kudo/nom: 57a6c7d | moritz++ | tools/build/Makefile.in:
[build] include Test.pir in the default make target
 15:52
kudo/nom: d248aa4 | jonathan++ | src/Perl6/Actions.pm:
Fix calls to post-declared listops. Means diag call in Test.pm works now.
 15:57
rrot: 8a9654e | NotFound++ | runtime/parrot/library/Test/More.pir:
a few more set_addr changed to set_label, TT #1857
 16:01
16:11 bubaflub left
dalek kudo/nom: b1eb40e | (Martin Berends)++ | t/spectest.data:
[t/spectest.data] comment out everything but the one script that passes
 16:13
kudo/nom: 41ae351 | pmichaud++ | src/core/ (3 files):
Replace some non-latin1 characters with "\\x" equivalents, to prevent poor parsing speed of CORE.setting.
 16:21
sella: c8a7614 | Whiteknight++ | s (5 files):
Add in two new tokenizer types for doing delimiter-based tokenizing
 16:24
sella: 83e629b | Whiteknight++ | src/unstable/template/ (6 files):
Add in first rough-draft of a templating library
sella: 6451e61 | Whiteknight++ | s (8 files):
Add Template library to the build. Fixes so things build
sella: 6633a17 | Whiteknight++ | s (5 files):
Fix several things so that the Literal, Eval, and Data node types work on toy test cases
sella: 1634fde | Whiteknight++ | src/ (3 files):
A few more additions and fixes to get the template library working in a general way
sella: bfd775e | Whiteknight++ | t/memoize/Memoize.t:
Merge branch 'master' of github.com:Whiteknight/Rosella
rrot: 796766d | Whiteknight++ | src/pmc/lexinfo.pmc:
Remove some unnecessary details from LexInfo. Hash.init_int sets the destroy and mark flags, so no need to re-set them. Don't do the gymnastics for making LexInfo const, the flag is worthless now
 16:26
rrot: 1f8dc33 | Whiteknight++ | src/ (2 files):
+revert an earlier commit which broke Lua. fperrad++
sella: bb79deb | Whiteknight++ | VERSION:
Bump Event version to 3. Add template as 0 Alpha
 16:27
kudo/nom: c54ff34 | jonathan++ | t/spectest.data:
Uncomment a test we pass.
 16:29
TT #1034 closed by NotFound++: attempt to access code outside of current code segment 16:37
TT #1034: trac.parrot.org/parrot/ticket/1034
16:49 JimmyZ left 17:05 rohit_nsit08 joined
dalek kudo/nom: f21cf10 | jonathan++ | src/core/Bool.pm:
prefix:<not>
 17:38
kudo/nom: 13e487b | jonathan++ | src/core/control.pm:
eval, with outer lexicals visible. Means eval_lives_ok and eval_dies_ok used in tests should work again.
17:43 cotto joined
cotto NotFound, ping 17:44
dukeleto TimToady sighting in the YAPC::NA hotel lounge 17:45
NotFound cotto: pong 17:46
cotto NotFound, how hard would it be to get winxed to generate something other than PIR?
specifically, M0 17:47
NotFound cotto: I have a project that generates C++
dukeleto NotFound: also, is there anything like a design document for Winxed?
NotFound: where does that code live?
cotto NotFound, a winxed that generates C++?
NotFound cotto: yes, but it doesn't handle many features yet. 17:48
github.com/NotFound/winxedxx
The main pitfalls is that an alternative generator must be highly coupled with the internals of the main compiler. 17:49
dukeleto NotFound: why so/ 17:50
NotFound dukeleto: because the compiler has not enough abstractions. 17:51
cotto goes to check in
17:54 bubaflub joined
bubaflub ~ 17:56
18:01 cotto left 18:13 cotto joined 18:38 rurban_ joined 18:41 rurban left, rurban_ is now known as rurban
lucian cotto: i'll consider downscaling the puffins. but they're so cute! 18:45
sorear pmichaud: does Rakudo have a snowball's chance of being able to use non-Parrot-bundled GMP bindings? 18:53
dalek kudo/nom: 75e02f2 | (Martin Berends)++ | t/spectest.data:
[t/spectest.data] mining spectest_full wins another 337 tests in 41 files
tadzik ah, spectest_full it is? 18:58
oh, ww
dalek kudo/nom: c1d7529 | moritz++ | src/core/IO.pm:
exit()
 19:01
19:11 ambs left
dalek nxed: 479cb0a | NotFound++ | winxedst1.winxed:
refactor a bit get_hll_global emision
 19:21
19:22 fperrad left 19:38 PacoLinux left
dalek kudo/nom: a104026 | jonathan++ | src/Perl6/Actions.pm:
Fix issue with re-stubbing something we stubbed already in BOOTSTRAP (should never happen in real Perl 6, mind).
 19:50
pmichaud 18:53 <sorear> pmichaud: does Rakudo have a snowball's chance of being able to use non-Parrot-bundled GMP bindings? 19:54
Yes.
NQP may very well do its own C-level bindings (for gmp as well as other stuff)
20:00 PacoLinux joined
dalek kudo/nom: 998c18a | pmichaud++ | src/ (3 files):
Move &exit into src/core/control.pm, add a nqp:: opcode for it.
 20:02
kudo/nom: 92fa053 | pmichaud++ | src/Perl6/Actions.pm:
Merge branch 'nom' of github.com:rakudo/rakudo into nom
20:05 ambs joined 20:10 dodathome left 20:11 dodathome joined, whiteknight left 20:21 PacoLinux left 20:29 mj41 joined 20:38 rohit_nsit08 left, dodathome left
dalek kudo/nom: 4e224d9 | jonathan++ | src/Perl6/Metamodel/MultiMethodContainer.pm:
Avoid a really obscure NQP bug. Nesting packages inside roles may well just be something we want to outlaw anyway, though...
 20:52
20:53 rohit_nsit08 joined 20:57 PacoLinux joined 21:19 ambs left 21:42 Psyche^ joined 21:47 Patterner left, Psyche^ is now known as Patterner 21:57 perlite_ joined 22:00 perlite left, perlite_ is now known as perlite 22:02 PacoLinux left 22:08 mj41 left
dalek nxed: 5b555bd | NotFound++ | winxedst1.winxed:
avoid creating a new closure when there are no lexicals
 22:10
kudo/nom: ace0918 | jonathan++ | src/ (6 files):
Fix up attributive parameter finding somewhat.
 22:21
kudo/nom: 9d200e0 | jonathan++ | src/core/EnumMap.pm:
EnumMap.Bool, EnumMap.elems.
kudo/nom: 49a6622 | jonathan++ | t/spectest.data:
Another passing test file.
 22:24
kudo/nom: 813e46c | jonathan++ | src/ (2 files):
Fix my Dog $fido .= new() style things, which show up in quite a few spectests.
 22:44
nxed: b4ecb18 | NotFound++ | t/ (3 files):
reorganize and improve some int operators tests
 22:45
22:47 bluescreen joined 22:49 lichtkind joined
dalek nxed: 3cbb4c2 | NotFound++ | t/ (2 files):
test exception type of div by zero
 22:51
kudo/nom: 6a3bb07 | pmichaud++ | src/ (2 files):
Restore whatever currying.
 23:02
23:04 mikehh left 23:08 mikehh joined 23:14 daniel-s joined
cotto ~~ 23:19
mikehh yo cotto, how you doin' 23:20
cotto mikehh, glad to be settled at yapc 23:21
dalek kudo/nom: 6885197 | jonathan++ | src/Perl6/Actions.pm:
our and my scoped methods.
kudo/nom: c55e033 | jonathan++ | src/core/traits.pm:
Stub in returns trait.
kudo/nom: 91c3937 | jonathan++ | t/spectest.data:
Two more passing tests.
mikehh cotto: anything interesting yet? 23:22
23:22 JaphMalik joined
cotto mikehh, it hasn't started yet 23:23
23:23 mikehh is now known as co, co is now known as mikehh
dalek kudo/nom: f0b4aea | jonathan++ | / (2 files):
Correct Code.ACCEPTS sig, winning another test file.
 23:24
mikehh bah typing in the wrong area
cotto: yes, but met anyone interesting or interesting discussions yet 23:25
23:27 lucian left
dalek kudo/nom: dc3f6cf | pmichaud++ | Test.pm:
Make sure isa_ok properly reports the type object.
 23:39
kudo/nom: 4f6bb1c | pmichaud++ | src/core/Array.pm:
Add Array.new().
kudo/nom: 16c2f0d | pmichaud++ | src/ (2 files):
Clean up for -> map translation, add Parcel.item.
kudo/nom: e6459bc | jonathan++ | src/Perl6/Actions.pm:
Fix setting of PAST::Block.arity so if $foo -> $x { ... } style things work.
 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!