Devel: 0.6.1 | parrotcode.org/ | 17/673 new/open tix
Set by moderator on 19 May 2008.
00:27 wknight8111 joined 00:33 kid51 joined 00:39 tedkat joined
DietCoke TimToady: I'm staring at that email trying to see what you changed. 01:01
TimToady s/a/e/ 01:05
01:05 cjfields joined
DietCoke ... wow. really. I even went into view raw source and I failed to see that. =-) 01:05
thank you. =-)
01:09 Theory joined
jhorwitz pmichaud: looks like trapping the exception willl work. should've thought of that before! 01:24
01:33 Psyche^ joined 02:07 teknomunk joined
DietCoke gets a successfull test run on freebsd. 02:31
02:48 particl1 joined 02:49 particl1 left 02:56 rdice joined 03:16 particl1 joined 04:01 barney joined
dalek r27654 | chromatic++ | trunk: 04:09
: [distro] NEWS updated; languages need more updates from language maintainers.
diff: www.parrotvm.org/svn/parrot/revision?rev=27654
04:20 particl1 left 04:22 grim_fandango joined
dalek r27655 | bernhard++ | trunk: 04:31
: More detailed NEWS on Eclectus and Plumhead
diff: www.parrotvm.org/svn/parrot/revision?rev=27655
04:56 Psyche^ joined
dalek r27656 | chromatic++ | trunk: 05:03
: [JIT] Temporarily disabled JIT for div_n_nc and div_i_ic ops; until PIR
: optimizer throws an exception on these ops when dividing by zero (or they get
: JITted to do the right thing then), they'll cause floating point exceptions.
diff: www.parrotvm.org/svn/parrot/revision?rev=27656
05:31 Zaba_ joined 05:37 bacek_ joined
nopaste "cognominal" at 82.67.232.89 pasted "problem with unicode" (47 lines) at nopaste.snit.ch/13020 06:13
06:16 uniejo joined
cognominal This is perl, v5.8.8 built for darwin-thread-multi-2level btw 06:16
06:30 teknomunk_ joined 06:45 ank joined 06:48 iblechbot joined
dalek r27657 | fperrad++ | trunk: 06:53
: [Lua]
: - fix test with luac2pir.pir
diff: www.parrotvm.org/svn/parrot/revision?rev=27657
r27658 | fperrad++ | trunk:
: [Lua]
: - bytecode translation : minor refactor
diff: www.parrotvm.org/svn/parrot/revision?rev=27658
r27659 | fperrad++ | trunk: 07:06
: More detailed NEWS on Lua
diff: www.parrotvm.org/svn/parrot/revision?rev=27659
07:26 integral joined 07:28 mj41 joined 07:56 zarchne joined, dcat joined, ewilhelm joined, obra joined, cotto-work joined, DietCoke joined, Ademan joined, avar joined, Piper joined, bacek joined, rgrjr joined, contingencyplan joined, particle joined, TimToady joined, tewk joined, cxreg joined, rhr joined, cout joined, Hinrik joined, shamu joined, cotto_home joined, workbench joined, GeJ joined, zev joined, toft joined 08:38 Zaba joined
dalek r27660 | fperrad++ | trunk: 08:53
: [PLATFORMS]
: - update for MinGW32
diff: www.parrotvm.org/svn/parrot/revision?rev=27660
09:15 IllvilJa joined
bacek hi there 09:36
I've investigated 'cmp bug'. 09:39
Actually, when we creates variables of subclass of Integer they are just Objects inside of parrot. 09:40
So, PMC_get_int doesn't work as supposed in src/pmc/integer.pmc. 09:41
On replacing PMC_get_int with VTABLE_get_integer everything works.
jonathan bacek: Sounds like a correct fix. 09:42
bacek But it looks like another bug: we actually call Parrot_Integer_cmp not on Integer class.
jonathan No, that is the cmp vtable method of the Integer class.
bacek But should be SELF be Integer?
s/be// 09:43
jonathan Yeah, in fact maybe you should be writing SELF.get_integer, which will produce a call through real_self, but all of that lot is meant to eventually go away. 09:44
bacek jonathan, ok. I'll create patch for src/pmc/integer which will call VTABLE_get_integer instead of PMC_get_int. 09:45
But it will slow down computations.
jonathan Sure, that's what abstractions do. :-) 09:47
bacek :) 09:48
jonathan We may just write out own Int PMC for Rakudo in the end or something.
We sorta-ish need it to inherit from Any, is the thing.
bacek Ho. actually all other methods uses SELF.get_integer()
So looks like this is "right way" :) 09:49
jonathan Yeah, I think that dispatches by real self (as in, to the Object, not just to the Integer), so it sees any overrides.
bacek jonathan, indeed 09:50
what about methods like set_integer_native? Should whey use virtual dispatch? 09:51
jonathan Does that actually dispatch? 09:52
I'd have thought that just set the value in the struct?
Oh, do you mean calls to set_integer_native rather than the method itself?
bacek jonathan, yes
jonathan Oh, yes, calls should be virtual, otherwise things go horribly wrong for subclasses. 09:53
bacek jonathan, got it.
jonathan Well, any subclass that overrides that method and expects to have it called anyway.
bacek jonathan, ok. So all 'getters' except in get_integer should call SELF.get_integer() instead of PMC_int_val. 'setters' still have to use PMC_int_val. Correct? 09:56
jonathan Hmmm. I'd have expected to be able to override set_integer as well, perhaps... 10:01
What 'setters' are you thinking of? 10:02
bacek VTABLE void thaw(visit_info *info) {
IMAGE_IO * const io = info->image_io;
SUPER(info);
if (info->extra_flags == EXTRA_IS_NULL)
PMC_int_val(SELF) = VTABLE_shift_integer(INTERP, io);
or it should be changed to SELF.set_integer(...)? 10:03
jonathan I'd leave thaw as it is.
Here we're trying to deserialize and just put the data back in place.
bacek jonathan, there is total 12 places where PMC_int_val is in use (after my changes). And I have no idea should I change them or not... 10:05
jonathan I suggest, send in a patch with the ones you're sure about, and mention those you aren't and ask for opinions. 10:08
bacek pastebin.org/37243
Looks like it is long mailqueue on RT...
moritz 12:08 < cognominal_> rakudo: my @a = (1); say @a[0]; 10:09
12:08 < exp_evalbot> OUTPUT[get_pmc_keyed() not implemented in class 'Integer'#current instr.: '_block11' pc 43
that's quite nasty :/ 10:10
cognominal I filed a bug and pmichaud is aware of it
bacek I'm 90% sure in patch on pastebin. I'll create another one for "im-not-definetly-that-this-is-right-thing"
moritz cognominal: ok
cognominal this is these small things known by the implementors but drives the beginner crazy. 10:11
jonathan They are also the things that look like they should be trivial to fix, but are a lot more subtle than you'd think. 10:13
bacek jonathan, exactly... 10:15
cognominal jonathan, I know because I tried :) 10:17
and my try was messy and wrong. 10:18
jonathan++ and pmichaud++ to explain me my errors 10:19
bacek jonathan, ok. There is methods: set_pmc, set_integer_same, set_integer_native, increment, decrement, thaw which uses PMC_int_val. 10:34
Plus init and initialize
And RT#46627...
So, should be MMD_Integer_EXACT used for check or this ticket outdated? 10:37
10:39 davidfetter joined 10:44 kid51 joined 10:48 ruoso joined
bacek jonathan, ? 10:50
jonathan sorry sorry, $DAYJOB is distracting me... 10:51
Got a link to that RT ticket?
BTW, method list that touch the value directly seem reasonable to me. 10:52
bacek rt.perl.org/rt3/Ticket/Display.html?id=46627 10:53
In src/pmc/integer.pmc:add() there is the todo item:
* TODO
* check for exact Integer type
* e.g. MMD_Integer_EXACT
jonathan, It's almost 9pm in Sydney :)
so my $dayjob already done
jonathan Ah, about 1pm here in Slovakia, so it's very much in progress here :-) 10:55
bacek BTW, Is different contexts implemented in rakudo? like ~@list?
jonathan I know +@list works
bacek jonathan, hey. It's lunch time!
jonathan Oh yeah!
Hmmm...I guess it depends if EXACT means it really is this very same class, or if it could be a subclass. 10:56
Not 100% sure on that, sorry.
I suggest either investigate, or comment on the ticket (cc'd to list) to get an answer. 10:57
bacek Ok.
rt.perl.org/rt3/Ticket/Display.html?id=54474 'cmp bug' with proposed patch 11:02
11:03 rdice joined
dalek r27661 | allison++ | pdd25cx: 11:22
: [pdd25cx] Bringing the pdd25cx branch up-to-date with trunk r27660.
diff: www.parrotvm.org/svn/parrot/revision?rev=27661
11:34 ambs joined
ambs \\o/ 11:34
DietCoke++ #finally answered my email 11:37
cognominal (positive reenforcement)++ 11:39
ambs :D
cognominal we are all pavlovian programmers 11:41
cognominal sweeps his keyboard
11:43 iblechbot joined
jonathan cognominal: I've just over lunch thought up a temporary and small hack to make "my @a = (1); say @a[0];" work...will give it a go later on today. Gotta sort out $dayjob stuff right now... 11:47
cognominal jonathan, check your mail 11:48
it seems there is a possibility of double room
jonathan Ah, OK 11:49
I'll take whatever I can get to sleep on! :-)
When I tried to book online I got a "no rooms available"...
cognominal my hack works if you leave out the "( )" case, second parameter of the action
I don't say it's clean 11:50
12:15 tewk joined 12:18 AndyA joined 12:19 Psyche^ joined 12:33 tetragon joined
bacek start loving assembler again. 12:56
rt.perl.org/rt3/Ticket/Display.html?id=54514
Implementing mergesort on such level is old good fun. 12:57
Like hacking soviet clone of PDP-11 back in 80-s :)
DietCoke bacek: any reason why you're redoing it in PIR when there's a version in C already? 12:58
(I ask because I just threw out language/tcl's PIR version of sort in favor of the C version. =-)
bacek DietCoke, It's fun! :)
DietCoke I cannot argue with that. =-) 12:59
bacek DietCoke, and its mergesort. Not quicksort. 13:00
DietCoke If you want to see tcl's usage, check out languages/tcl/runtime/builtin/lsort.pir , look for 'sort' with the quotes. (I'm passing in a comparator function that's written in PIR)
moritz bacek++ # -Ofun
DietCoke but otherwise carry on.
13:04 Khisanth joined, spinclad joined, Tene joined, zostay joined, TonyC joined, nopaste joined, diakopter joined, purl joined, japhb joined, silug joined, Eevee joined, Sartak joined, dngor joined, jjuran joined, jjore joined, confound joined, Infinoid joined
bacek DietCoke, hmm... You invoke $P0.'sort'. 13:04
DietCoke, how it call to Parrot_quicksort.
I just can't find it...
pmichaud bacek: the ResizablePMCArray PMC already knows how to 'sort'
i.e., doing $P0.'sort' on a ResizablePMCArray ultimately calls Parrot_quicksort 13:05
src/pmc/fixedpmcarray.pmc, lines 41-46 13:06
bacek > ('a',1,'b',2).sort.say
Method 'sort' not found for invocant of class 'List'
its in rakudo...
pmichaud well, one would have to pass a comparison function to sort
it doesn't have a default
still, I'm surprised that it gives "Method 'sort' not found" -- just a sec 13:07
bacek pmichaud, ok. i'll try it
pmichaud very interesting. 13:09
purl it has been said that very interesting. is that a quote?
pmichaud Apparently the 'sort' method doesn't get inherited by subclasses of ResizablePMCArray . 13:10
(such as 'List')
bacek pmichaud, heh...
nopaste "pmichaud" at 76.183.97.54 pasted "'sort' not inherited by subclasses of RPA" (27 lines) at nopaste.snit.ch/13022 13:11
bacek pmichaud, Can't find 'sort' in "Resizable"... But can in 'Fixed'...
pmichaud ResizablePMCArray isa FixedPMCArray 13:12
see line 25 of resizablepmcarray.pmc
pmclass ResizablePMCArray extends FixedPMCArray need_ext provides array {
bacek pmichaud, a! Got it...
pmichaud does a double-take 13:13
oh this is totally bizarre
13:13 Zaba joined
pmichaud big bug :-) 13:13
particle did we make the conversion from DYNSELF to SELF? 13:14
bacek hides
particle it may be that you need to say DYNSELF.some_vtable_function for inheritance to work
nopaste "pmichaud" at 76.183.97.54 pasted "'sort' not inherited by subclasses of RPA, but 'push' is" (27 lines) at nopaste.snit.ch/13023
particle part of the pmc pdd is to convert DYNSELF to SELF and SELF to STATICSELF
pmichaud I'm guessing that superclass methods aren't being inherited properly 13:15
i.e., given that RPA is a PMC class and RPA inherits from FPA
if we create a subclass of RPA, then the methods of FPA aren't being included in the subclass 13:16
particle perhaps.
you can test by creating a subclass of FPA
see if sort works
pmichaud oh, good idea.
purl pmichaud: Good Idea: Visiting the circus. Bad Idea: Having the circus visit you.
pmichaud ah, it's more complex than that. 13:18
bacek: short answer: until we get inheritance working, we need a 'sort' method for rakudo :-) 13:19
bacek: however, you could copy all of the elements of a list into a FixedPMCArray, and then use its sort method
bacek pmichaud, oooo... Good answer :) 13:20
pmichaud and then cast that back into a List
particle that won't work with lazy lists, but we don't implement those yet
well, i suppose it will work... just need to make it unlazy before conversion to FPA 13:21
pmichaud correct
bacek ok. Is there anything useful that "casual developer" can implement for rakudo?
pmichaud hard to sort a lazy list :-)
except in special cases
bacek pmichaud, why? Just use old good bubble-sort
pmichaud well, we can sort a lazy list, but it's no longer lazy. 13:22
particle bacek: look for functions from S29 that aren't implemented
bacek particle, 'sort' for example? :)
pmichaud also, note that the 'sort' method is an inplace sort
(in FixedPMCArray)
let me finish updating NEWS, then I'll file a ticket for 'sort'
particle pmichaud: can we implement gather/take yet? 13:23
particle updates and checks roadmap
pmichaud particle: I haven't really looked at it yet. That would be way down the list on roadmap, though
moritz you need some sort of dynamic scoping for gather/take
even if it's just global variable in the interpreter of some sort 13:24
particle that we can do
pmichaud oh, dynamic scoping isn't too hard
that's pretty easy in Parrot.
bacek start thinking that implementing sort as creating FPA is good idea...
13:24 Khisanth joined, spinclad joined, Tene joined, zostay joined, TonyC joined, nopaste joined, diakopter joined, purl joined, japhb joined, silug joined, Eevee joined, Sartak joined, dngor joined, jjuran joined, jjore joined, confound joined, Infinoid joined
moritz actually sort() could return a lazy list 13:25
pmichaud bacek: for now I'd be happy with any 'sort' implementation that works, whether it uses the one in FPA or not.
moritz than it could run in O(n) if you only take the first or first two items
and O(n log n) if you take them all
pmichaud well, taking only the first item is effectively "min" :-) 13:26
but yes, it would be possible for sort to return a lazy list. 13:27
particle yep, and O(n) may be very large for list items calculated from values stored in a db
dalek r27662 | allison++ | pdd25cx:
: [pdd25cx] Adding new exception type.
diff: www.parrotvm.org/svn/parrot/revision?rev=27662
r27663 | pmichaud++ | trunk: 13:29
: Update NEWS.
diff: www.parrotvm.org/svn/parrot/revision?rev=27663
13:31 rdice joined
DietCoke pmichaud: (doesn't work) but it works in tcl... odd. 13:33
pmichaud does tcl subclass ResizablePMCArray ?
DietCoke checking.
yes.
pmichaud is the subclass a PIR subclass or a PMC subclass ? 13:34
DietCoke PMC
pmichaud ah, that's the difference.
DietCoke ISTR there's a ticket about the PIR from PMC not quite working.
pmichaud yes, but I thought that had been resolved long ago.
but it's not working for superclasses of PMC classes
DietCoke possibly, but if you're saying it doesn't work...
pmichaud I have a simple test case written -- just cleaning it up 13:35
DietCoke how do you superclass a PMC? =-)
(test) hokay.
pmichaud (it's not working for methods of superclasses of PMC classes...)
nopaste "pmichaud" at 76.183.97.54 pasted "methods in PMC classes and inheritance - b0rken" (39 lines) at nopaste.snit.ch/13024 13:38
jonathan Segfault bad. 13:39
pmichaud hmm. 13:40
in this case it's because 'lower'() does bad things.
PMC * const ret = pmc_new_noinit(INTERP, SELF->vtable->base_type);
PMC_str_val(ret) = s;
particle excellent test case, pmichaud 13:41
pmichaud so, segfault here is due to the 'lower' implementation, not inheritance. Let's try...
jonathan PMC inheritance isn't working yet as PDD 17 spec'd it, I don't think. 13:42
nopaste "pmichaud" at 76.183.97.54 pasted "inheritance b0rken in subclasses of PMCs -- #2" (32 lines) at nopaste.snit.ch/13025
pmichaud well, it works for at least one level.
for example, the 'push', 'pop', etc. methods defined in RPA work for PIR subclasses of RPA 13:43
bacek pmichaud, my version of 'sort' attached to ticket.
pmichaud, and i try to use FPA...
jonathan pmichaud: I meant more that, the way it's implemented now. 13:45
pmichaud here we go -- this illustrates it nicely. 13:48
nopaste "pmichaud" at 76.183.97.54 pasted "inheritance b0rken in subclasses of PMCs -- #3" (43 lines) at nopaste.snit.ch/13026
pmichaud filing ticket(s) 13:49
bacek pmichaud, looks like I found another bug... 13:50
bacek@icebolt:~/src/parrot/languages/perl6$ cat s3.t 13:51
my @l = ('a', 'b', 'c', 'd', 1, 2, 3, 4);
my @l1 = sort { $^b <=> $^a } , @l;
say ~@l1;
bacek@icebolt:~/src/parrot/languages/perl6$ ../../parrot perl6.pbc s3.t
DietCoke ponders asking ambs to post his script.
bacek It uses FPA sorting...
13:52 Andy joined
bacek pastebin.org/37282 impl. 13:52
dalek r27664 | coke++ | trunk: 13:56
: update freebsd
diff: www.parrotvm.org/svn/parrot/revision?rev=27664
r27665 | allison++ | pdd25cx: 14:06
: [pdd25cx] Changing name of 'dumpcore' macro to 'DUMPCORE' to match coding
: standards, with additional comment for clarity.
diff: www.parrotvm.org/svn/parrot/revision?rev=27665
14:07 jhorwitz joined
dalek r27666 | allison++ | pdd25cx: 14:09
: [pdd25cx] Missed one DUMPCORE.
diff: www.parrotvm.org/svn/parrot/revision?rev=27666
14:10 wknight8111 joined
dalek allison@perl.org | Concurrency Tasks: 14:11
link: www.perlfoundation.org/parrot/index...ency_tasks
shorten dalek's url is at xrl.us/bjqpy
jhorwitz pmichaud: ping
bacek $ cat s3.t 14:16
my @l = ('a', 'b', 'c', 'd', 1, 2, 3, 4);
my @l1 = @l.sort;
say ~@l1;
$ ../../parrot perl6.pbc s3.t
1 a 2 b c 3 4 d
yeeek...
particle who's sort is that? 14:17
pmichaud jhorwitz: pong (but has to be quick -- have to go to a lawyer's office in 13)
bacek FPA 14:18
pmichaud (er, have to *be* at a lawyer's office in 13)
(sort): what comparison function is being used?
bacek pastebin.org/37282
infix:cmp
get_global $P0, "infix:cmp"
sorted = 'sort'($P0, list :flat)
pmichaud oh. perhaps infix:leg would be a better choice, since we know that parrot's 'cmp' isn't good.
jhorwitz pmichaud: catching the exception worked 14:19
bacek pmichaud, same results
pmichaud jhorwitz++ # excellent
jhorwitz all tets pass, should i commit or wait for after the release?
tests
pmichaud I think committing is okay. Or it can wait for the release -- your choice.
particle be careful with core parrot changes 14:20
language changes are ok
pmichaud TimToady++ helped get jonathan and I set aright on how to handle the $!, $/, $! items
jhorwitz particle: this is for rakudo
particle ship it!
jhorwitz giddyup
pmichaud language changes are okay, yes, but we'd still like to make sure that rakudo passes 'make test' in a release.
bacek voting for rt.perl.org/rt3/Ticket/Display.html?id=54514
jonathan Aye, I may well look at that tomorrow
($1, etc, not the ticket just pointed at...)
particle pmichaud: yes, and i'll run make test :)
jonathan erm, $! 14:21
pmichaud gotta run to lawyers -- bbiaw
jonathan bacek++ # sort patch 14:22
dalek allison@perl.org | Concurrency Tasks:
link: www.perlfoundation.org/parrot/index...ency_tasks
shorten dalek's url is at xrl.us/bjqpy
bacek and rt.perl.org/rt3/Ticket/Display.html?id=54474 14:23
oops. 14:24
14:24 Psyche^ joined
bacek compiler complains about 54474 14:24
14:27 davidfetter joined
Tene Okay, I'd really like to figure out all the issues with properly handling $_ as a block parameter. 14:30
Maybe I should send mail to the list. 14:31
particle did you see timtoady, jonathan, and pmichaud talking last night about $_?
the answers lie within
Tene Fantastic! 14:32
14:32 NotFound joined
Tene Last night was perhaps the wrong night to take a break. 14:32
NotFound Hello.
Tene particle++ # information!
Tene scrollback 14:33
purl scrollback is your friend
dalek r27667 | jhorwitz++ | trunk: 14:36
: catch 'No such caller depth' when subs are called from embedded environment (RT #54444)
diff: www.parrotvm.org/svn/parrot/revision?rev=27667
14:38 radhios joined
bacek 27666 crashes on make test... 14:39
Tene Okay, that's some good reading for lunch. 14:45
bacek Failed 5/584 test programs. 19/11275 subtests failed.
bacek looser...
purl it has been said that looser is an adjective
bacek please anyone check patch in rt.perl.org/rt3/Ticket/Display.html?id=54474. I'm not sure about correctness... 14:48
Tene looks. 14:52
integer2.diff?
bacek Tene, yep 14:53
Tene Does it cause any new test failures? 14:54
I don't see any problems with it. 14:55
bacek Tene, I'm running tests right now... It takes a long time...
Tene nods.
Hm. I've got an eight-cpu box at my office... I should run tests there.
Tene bbiab
bacek I run them on my old laptop... 10 minutes... I... can't... wait... any... more... 14:56
$ prove t/pmc/integer.t 14:57
t/pmc/integer......ok
All tests successful.
purl Are you feeling lucky?
dalek r27668 | allison++ | pdd25cx:
: [pdd25cx] Change 'push_new_c_exception_handler' to 'Parrot_ex_add_c_handler',
: and remove 'new_c_exception_handler'.
diff: www.parrotvm.org/svn/parrot/revision?rev=27668
bacek My middle name is 'bad'... 14:58
Tene bacek: looks okay to me, then.
bacek purl, whanna cookie?
purl i don't know, bacek
bacek Tene, it's good. 14:59
purl, cookie?
purl C is for cookie, it's good enough for me!
dalek allison@perl.org | Concurrency Tasks:
link: www.perlfoundation.org/parrot/index...ency_tasks
shorten dalek's url is at xrl.us/bjqpy
Tene purl: no, C is for lettuce, it's good enough for me!
purl okay, Tene.
Tene Hm, that was the wrong factoid.
Oops.
bacek purl, say squaak 15:00
purl bacek: i'm not following you...
bacek Tene, same number of failing tests. 15:03
Files=584, Tests=11275, 501 wallclock secs...
ok. time to sleep. 15:10
15:21 Theory joined
pmichaud jhorwitz: ping 15:24
jhorwitz pmichaud: pong
pmichaud jhorwitz: can we close #54444 ?
jhorwitz i was thinking about that. i guess we can, unless we want to question the behavior of $P0['lexpad';1]
that could be a separate ticket though
pmichaud I think it's reasonable to question $P0['lexpad';1], but that probably belongs in the parrot queue. 15:25
so yes, it'd be a separate ticket.
jhorwitz the let's close it out! :)
pmichaud and it may be a moot point anyway, given how $! $/ $_ are to be handled so I think we just close it.
jhorwitz sounds good to me.
pmichaud likes seeing tickets get closed.
davidfetter mmm...closed tickets :d 15:26
jhorwitz likes 24 hours turnarounds :)
pmichaud it will be an interesting milestone when rakudo has more outstanding tickets than parrot :-)
davidfetter figures outstandingness is in the eye of the beholder
particle pmichaud: 'say False' doesn't work. 15:27
Bool::False, sure.
pmichaud rt#54444 closed. 15:28
where is 'False' defined?
is it an imported global, or ... ?
particle src/classes/Bool.pir
purl src/classes/Bool.pir is where i was looking
pmichaud no, I mean in the spec.
particle ah. gotta check that. 15:29
pmichaud S12
our bit enum *bool <False True>;
our bit enum *taint <Untainted Tainted>;
particle hrmm. we don't do enum now.
or bit, for that matter
pmichaud for now, just create global entries for False and True that point to 0 and 1 15:30
particle so, for the spectests, should i modify them to use Bool::False except when explicitly testing False?
...or should i create globals in Bool.pir?
or both?
purl both are highly experimental at this point though
pmichaud create globals. 15:31
particle forget both
purl particle: I forgot both
particle ok.
pmichaud "ack False" shows that it occurs enough in the spec that we might as well create the global.
jonathan might be interested in implementing enums, since they're really just classes acting as roles :-) 15:32
jonathan Ooh! Shiny feature! 15:33
pmichaud but for now I think it's a reasonable workaround to create these as globals.
how often does False appear in the spectests? 15:34
particle i'm looking at S03-operators 15:35
haven't looked outside there yet
not.t and true.t, and i'll be adding some tests, and probably false.t
hrmm. maybe not. is there a 'false' op?
pmichaud S12 has:
particle acks spec 15:36
pmichaud Never compare a value to "C<true>", or even "C<True>". Just use it
in a boolean context.
so, I think that S03-operators is wrong to be using False here
if we're testing boolean-ness, we should be using ok() or nok()
particle ok, so that should be nok() yep 15:37
pmichaud if we're testing the type of the returned value, we need to use =:= or something similar
or check the .WHAT() value, or something like that.
(and =:= doesn't quite work in rakudo yet)
particle may i check that Bool::True != Bool::False? 15:38
pmichaud sure
particle or that True == Bool::True?
pmichaud keeping in mind that True == Bool::True is really the same as +(True) == +(Bool::True)
i.e., it's a numeric comparison, not a value comparison. 15:39
particle yep
pmichaud I think more useful is to ask Bool::True == 1
particle in not.t, i want to test explicitly that True != False, etc
we can't guarantee that Bool::True == 1 can we?
it may depend on the vm/compiler/platform's identity of truth 15:40
pmichaud it's a question of numification
particle ok, then it's +Bool::True == 1 15:41
pmichaud except that infix:<==> implies numeric context on its arguments
so Bool::True == 1
particle hrmm. ok. 15:42
pmichaud I can't find that explicitly in the spec, fwiw.
particle ah. === can't be tested, though
or =:=
pmichaud the spec does say that Order::Increase, Order::Same, Order::Decrease numerify to -1, 0, +1, so I think it's reasonable to expect Bool::False and Bool::True to numerify to 0, 1 15:44
Tene 1, 0, -1 == Bool::True, Bool::False, Bool::FileNotFound?
pmichaud heh 15:45
FileNotFound is likely an exception of some sort, though.
moritz lol
particle i'm kinda surprised Bool::True isn't a Failure
pmichaud Bool is, though.
dalek Michal Jurosz | RFP: Parrot Needs Better Smoke Reports: 15:48
link: www.perlfoundation.org/parrot/index...ke_reports
shorten dalek's url is at xrl.us/bfeof
pmichaud oh, I guess Bool is just an undefined value -- that's not necessarily a Failure. 15:52
15:56 cjfields joined
Tene Mmm. make -j 9 test 16:04
real\t3m6.804s 16:09
user\t1m51.388s
sys\t0m45.239s
particle hrmm, there's no nok() func in fudge or Test.pm 16:14
Tene Hm. 16:18
I wonder if rakudo makes a new block for this: print $_ for @a;
particle i expect that what's left of 'for' is a closure 16:21
purl I hardly KNOW her.
Tene Yeah. That makes sense.
17:07 <@TimToady> and the presence of $_ is not enough to force "I must have a parameter" 16:25
A bare closure without placeholder arguments that uses $_ (either explicitly or implicitly) is treated as though $_ were a formal parameter:
From S04
So I guess that statement *only* applies to bare closures.
Wait, formal parameter with a default of OUTER::<$_> 16:26
So do I need to figure out which blocks are "bare" and treat them differently? 16:27
particle i assume bare means not a sub or pointy block 16:28
Tene Ah.
So for, while, etc. are bare?
particle doesn't for take a pblock? 16:29
gotta check STD on that
Tene It can.
Tene sleeps. 16:30
erm... no.
Tene teaches.
particle just half your brain sleeps, then.
DietCoke allison: ping. 16:39
allison dietcoke: ya
16:40 braceta joined
DietCoke was going to comment on r27670, but I think I need to see the whole function. moment. 16:41
ssh to feather sloooow all of a sudden.
would you mind a small rework on that that instead of nesting the ifs, did each comparison in sequence, failing first rather than last? 16:44
allison sure
DietCoke also, does it matter if you check base_type -and- isa ?
doesn't base type win?
allison just the base type is good enough 16:45
DietCoke k.
allison (it was checking base_type || isa, but that was just wrong
also it was checking base type against a literal, which didn't allow for inheritance 16:46
16:46 Zaba joined
allison oh, do you want to make the change, or shall I 16:46
DietCoke either or.
allison go for it, I'm digging into the 'can_handle' bullet point (the last one aside from language testing) 16:47
16:48 Psyche^ joined
DietCoke is anyone logged into feather? according to ssh I'm connecting, but then hanging. 16:51
particle svn.pugscode.org is down, too 16:52
16:54 ambs joined 17:08 radhios joined
jhorwitz hm, getting "attempt to access code outside of current code segment" in my pure perl mod_perl6 code 17:19
jhorwitz figures out where to even start deubgging that 17:20
particle the joy of embedding parrot.... 17:21
jhorwitz it obviously works with the PIR version of mod_perl6
DietCoke that sounds suspiciously like the inferior runloop!
particle yes, it does.
purl if you say so...
jhorwitz inferior cuz it doesn't work? ;-)
DietCoke Sounded like some CS term. 17:22
me, being a lowly engineer, just nod my head.
particle inferior like the flyers.
jhorwitz that was low 17:23
particle oh, but it was fun :)
jhorwitz but all will be well when cindy is put in his place
17:27 AndyA joined
jhorwitz does each compile and load_bytecode produce a separate code segment? 17:30
particle i believe so 17:33
jhorwitz hm. then i don't understand how code from one segment can call a sub from a load_bytecode w/o generating that error. apparently there's lots i don't understand. :) 17:34
17:35 chromatic joined
particle oops, i'm probably getting confused with compilation unit 17:36
jhorwitz that would make sense.
particle chromatic will tell you whether each compile and load_bytecode produce a separate code segment, jhorwitz
jhorwitz this chromatic...where can i buy one? 17:37
chromatic chromatic store, where else? 17:38
jhorwitz we don't have those out east
chromatic I *believe* they produce separate packfile structures, which mean separate code segments.
17:38 braceta joined
chromatic Okay, the first part I know. The implication I believe. 17:38
particle remember when jonathan started the bytecode changes, and disappeared for a while? 17:44
seems the same thing happened to infinoid.
"don't touch the bytecode pdd."
jhorwitz points fingers 17:46
Infinoid oh yeah, pdd13, I still have a stack frame pointing to that, around here somewhere 17:48
DietCoke juerd? 17:49
purl juerd is root or at juerd.nl/
DietCoke juerd email?
17:50 pmichaud joined
particle juerd@ 17:51
juerd is also juerd@juerd.nl 17:52
purl okay, particle.
17:52 pmichaud joined
DietCoke emailed him. 17:53
pmichaud feather is incommunicado? 17:54
17:54 dalek joined
particle yes, so it seems 17:55
pmichaud oh, it just came back up.
particle here's dalek!
yay
pmichaud unless it's up-down-up-down-up-down.
particle well, i finally was able to commit to svn.pugscode.org
chromatic That's good. It was sick of you stringing it along. 17:56
pmichaud +34 to #parrotsketch
chromatic pmichaud, all tests passed using autoboxing for that lower method call you suggested.
How urgent is the fix? 17:57
pmichaud not
not urgent
Tene Ack, #ps in .5
17:57 jonathan joined
dalek r27673 | particle++ | trunk: 17:57
: [rakudo] add 'nok' sub to Test.pm
diff: www.parrotvm.org/svn/parrot/revision?rev=27673
Tene svn log | grep tene
pmichaud I only noticed it while trying to find out about inheritance among pmc types, so I wanted to ticket it before I forgot about it
but in scanning I noticed a number of places where methods are creating return PMCs for base type values and they no longer need to be (and are more robust if they don't) 17:58
chromatic Robust in the sense that PMC_x_val is not robust?
pmichaud that, plus also the fact that autoboxing honors .hll type mapping 17:59
and the fact that it's silly to put the result into a PMC if we're just going to cast it right back to an int
i.e., $S0 = $P0.'lower'() was creating an unnecessary PMC.
18:00 Juerd joined
chromatic Makes sense. 18:00
Tene pre-reports 18:03
I'm not sure if that's the right time interval. Didn't check the dates. Sounds right, though.
particle cia.vc/stats/author/tene 18:05
the last week has day of week, before that is listed by date 18:06
DietCoke mmm, reduce pmc pressure.
18:06 Ivatar joined
Tene Ack, I need to commit my List::uniq patch. 18:07
eh, I'll do it during the middle of the release.
DietCoke :P 18:08
chromatic If you want to commit now, go ahead.
18:16 braceta joined
dalek allison@perl.org | Concurrency Tasks: 18:17
link: www.perlfoundation.org/parrot/index...ency_tasks
shorten dalek's url is at xrl.us/bjqpy
18:18 wolverian joined 18:21 pmichaud joined 18:26 barney joined
DietCoke PS in 2 18:28
from juerd: 18:45
Yes, Apache2 consumed all memory again. I've stopped apache2, and am
hoping someone will volunteer to figure out what is wrong because I
don't have the time.
pmichaud how long is apache2 running before it eats up all memory? 18:47
Juerd A few days
pmichaud okay, looking.
Juerd If you need root access let me know. 18:48
pmichaud you might not know the answer, but is it the master process consuming all of the memory or one of the subprocesses?
Juerd I don't know.
I know that the linux kernel spits out messages about processes called apache2
See c14.convolution.nl/munin/perl6.nl/f...emory.html 18:49
shorten Juerd's url is at xrl.us/bkshd
pmichaud it looks to me as though apache2 is set with "MaxRequestsPerChild 0"
this means that child processes never exit
often that's okay, but it can also mean that memory consumption can grow big if something gets out of whack
Juerd This started a week ago
I set that to 100 18:51
pmichaud is apache2 heavily loaded on feather? If not, then setting a MaxRequestsPerChild value might be worthwhile to make sure that child processes are recycled
Juerd pmichaud: Not at all
dalek r27674 | allison++ | pdd25cx:
: [pdd25cx] Fix trailing whitespace.
diff: www.parrotvm.org/svn/parrot/revision?rev=27674
r27675 | chromatic++ | trunk:
: Updated PLATFORMS for Aix/PPC and Solaris 10/gcc/cc (Paco, RT #54524).
pmichaud yes, I'd recommend MaxRequestsPerChild then. It's not as if there's a huge cost to starting up an additional child every so often
dalek diff: www.parrotvm.org/svn/parrot/revision?rev=27675
r27676 | allison++ | pdd25cx:
: [pdd25cx] Change 'can_handle' in ExceptionHandler to allow for subclasses of Exception.
diff: www.parrotvm.org/svn/parrot/revision?rev=27676
r27677 | fperrad++ | trunk: 18:52
: [RELEASE]
: update LANGUAGES_STATUS.pod from languages-smoke results
diff: www.parrotvm.org/svn/parrot/revision?rev=27677
Juerd I've started apache2. To all: if you have any time in between, please run top and evaluate apache2's memory use
dalek r27678 | fperrad++ | trunk:
: [RELEASE]
: update LANGUAGES_STATUS.pod from manual tests
diff: www.parrotvm.org/svn/parrot/revision?rev=27678
DietCoke Juerd++
pmichaud Juerd: I'll check in on it every few hours
Juerd Great; thank you
pmichaud Andy: how do I submit an article to rakudo.org, btw? 19:00
19:04 Zaba_ joined
pmichaud chromatic: I think you should be good-to-go as far as CPAN indexing is concerned. 19:04
chromatic Thank you. 19:05
particle thanks for doing that, pmichaud++, i got distracted by a phone call about a job 19:10
pmichaud: i think the mt.cgi link is on perlbuzz to login to rakudo.org 19:12
dalek r27679 | chromatic++ | trunk:
: [distro] Updated scaffolding for 0.6.2 release.
pmichaud I try using the mt.cgi link, but it just cycles me back to the perlbuzz page.
dalek diff: www.parrotvm.org/svn/parrot/revision?rev=27679
particle perlbuzz.com/cgi-bin/mt/mt.cgi
well, i can't remember my password, so i'm no help 19:14
Eevee openid++ 19:15
dalek r27680 | chromatic++ | trunk:
: [distro] Added a missing line from NEWS.
diff: www.parrotvm.org/svn/parrot/revision?rev=27680
pmichaud whenever I follow the mt.cgi link, it asks for my username and password. I enter it, and it takes me to the "Your Profile" form. I can go back to the original (perlbuzz) page and I appear to be logged in, but I don't see any way to post a new article. 19:19
(and, of course, my goal is to contribute to rakudo.org, not necessarily perlbuzz)
particle yes, rakudo.org uses the same mt instance, so there's only one login iiuc 19:20
jonathan pmichaud: Maybe you have an accout, but haven't been given permissions to post on rakudo.org?
pmichaud that's my guess :-)
jonathan I can't grant you 'em, I already checked that.
particle bug andy 19:21
pmichaud 19:00 <pmichaud> Andy: how do I submit an article to rakudo.org, btw?
:-)
particle Andy ack alester perlbuzz splint lint headerizer petdance # all points bulletin :) 19:22
...if that doesn't make his irc client explode... 19:23
Andy Who's talkin' about me? 19:24
pmichaud ALL OF US! 19:25
purl I think that we're all bozos on this bus. Except of course for the bots.
Andy CAP'N! SHE CANNAE TAKE MANY MAR KEYWERDS!
pmichaud Andy: how do I submit an article to rakudo.org?
Andy Let me see if you have permsssions on there.
Have I given you permissions before?
pmichaud the group consensus is "probably not"
I thought you had, but perhaps we only talked about it. 19:26
particle needs a password reset, too
pmichaud someone suggested that I should cross-post some of my use.perl articles to rakudo.org
Andy oh yes please. 19:28
jonathan takes blame for the suggestion 19:29
pmichaud: I was lying in bed last night trying to sleep and failing to...and thinking about :scope('context')
I'm thinking maybe we should do something more general 19:30
pmichaud ...as in?
jonathan We're already going to need to have a way to cutomize context anyway
chromatic Release tagged, tarball made and tested, uploading now.
Andy ok, you're now a contributor on rakudo 19:31
dalek r27681 | chromatic++ | RELEASE_0_6_2:
: Tagged release 0.6.2 from r27680
Andy not sure I can reset your PW. LOoking.
dalek diff: www.parrotvm.org/svn/parrot/revision?rev=27681
19:31 PerlJam joined
jonathan :scope('custom'), and some other way to say how to do a get and store 19:31
DietCoke chromatic++
jonathan chromatic++
NotFound Someone can take a look at #54384 ?
confound chromatic++
(everyone else involved, too)++
Andy ok, I just reset it.
pmichaud oh. well, for :scope('custom') someone could actually do that by deriving a new class from PAST::Val and adding their custom scope as a method
Andy Password was reset for user 'pmichaud' (user #97). Password was sent to the following address: pmichaud@pobox.com 19:32
chromatic I think creating a new function of a similar name but different behavior is a mistake, NotFound.
pmichaud Andy: thanks!
chromatic I'm starting to wonder if disallowing NULL strings in bytecode is wise.
jonathan Yeah, true.
Andy pmichaud: Thank YOU.
Rakudo.org is happy to have much content.
pmichaud okay, I have to go pick up kids from school.
bbi30
jonathan It just seems to me like Perl 6 has very specific needs out of :scope('context')
ok, sure
NotFound chromatic: agree, but I was worried about breaking the pmc that uses the current version. 19:33
pmichaud (context) well, perhaps Rakudo should be the case that demonstrates adding custom PAST node types then :-)
I just know that I have enough use for context variables in implementing various tools that it has generic value
jonathan Ah, OK, that was going to be my question...is it generic enough. 19:34
pmichaud a lot of things in PGE and PCT can be made a lot simpler if there's a generic context capability available
chromatic One problem with adding NULL checks to ops (another idea) is that we have to do those checks in the JIT as well, and that's not trivial.
At least, I find modifying ops in the JIT non-trivial.
DietCoke Is it helpful to put in #line directives in the C generated from PMCs? more often than not it's getting in my way.
pmichaud I tend to agree that custom handling of null seems wrong.
okay, gotta run. back in a while.
particle andy: could you reset my pw, too?
Andy sheesh 19:35
NotFound chromatic: the check is not done at the op level, but at a function called by the opcode.
Andy ANYONE ELSE? HUH?
dalek r27682 | fperrad++ | trunk:
: [Lua]
: - bytecode translation : minor refactor
diff: www.parrotvm.org/svn/parrot/revision?rev=27682
Tene Andy: mine too!
chromatic I know. I'm just brainstorming different approaches.
particle chromatic++ andy++ surliness++
Andy Tene: Do you actually have one?
Tene Well, no. You used capital letters, though.
Andy Password was reset for user 'particle' (user #131). Password was sent to the following address: jerry.gay@gmail.com 19:36
NotFound chromatic: I've fixed some days ago a problem caused by that, the opcode function checked for null but the jitted version not.
Eevee Andy: me too! (I don't have an account but heckling you seems to be the popular thing to do)
particle thanks andy!
Andy You're welcome particle!
NotFound So doing it in the opcodes is at least error-prone. 19:37
Andy The rest of you can go jump in a lake!
Rrraaarrr!
chromatic How about a creek? I have a creek, not a lake.
Eevee I have a bathtub
PerlJam Andy: give me an account so that I too may ask that you reset my password. :)
and then we can all jump in a creek and splash around like school children 19:38
Eevee there are now more people without accounts asking than people with
jonathan Will if all of the people asking for them are actually going to post... ;-) 19:39
PerlJam I don't even know what accounts people are talking about, but if it's parrot or perl6 related and I have time, I'll try to post. 19:40
19:41 ambs joined
Eevee I'm not sure I have anything to say yet that others couldn't say better 19:43
NotFound Eevee: maybe the others can but doesn't. 19:45
PerlJam Eevee: Or maybe your perspective is just as valuable as anyone else's perspective (more so if you share). 19:46
NotFound And in the worst case we can have a lot of fun blaming you }:) 19:47
Eevee oh gosh so much pressure 19:48
barney quit
jonathan Eevee: If you're contributing stuff, you have a pretty good perscpective on the thing you're contributing. ;-)
pressure++ ;-)
japhb Cue Queen 19:49
NotFound Press any Eevee to continue.
Eevee they told me about this in middle school 19:50
I'm supposed to just say no
jonathan: when I manage to make it past one patch I will perhaps consider myself 'contributing' :P 19:51
chromatic /topic #parrot Parrot 0.6.2 "Reverse Sublimation" Released | parrotcode.org/ | 19/674 new/open tix 19:52
Eevee but! I will keep it in mind
moderator Parrot 0.6.2 "Reverse Sublimation" Released | parrotcode.org/ | 19/674 new/open tix 19:52
Eevee may want to update parrotcode.org 19:53
japhb YAY!
DietCoke Eevee: one thing at a time.
chromatic wishes crow.pir added the next release date. 19:55
particle hrmm. that'd take some doing :( 19:56
unless it's blind lookup in a file
or somebody writes DateTime.pir
ambs \\o/ 19:57
chromatic It's in the JSON file. 19:58
particle oh. well then, that's *simple*
chromatic ... or at least a placeholder, 'cuz I forgot it.
particle i don't remember ever announcing the next release date for a release 19:59
pmichaud (contributions) from a marketing perspective, what makes more people likely to participate is if they see other people of about the same perceived skill level participating
chromatic step nine: Don’t forget to include the next scheduled release date.
pmichaud so posts from anyone who is working/playing with Parrot/Rakudo are helpful, even if they seem "trivial" 20:00
so, I can start ripping out deprecated features now? 20:01
pmichaud clicks his Code Ripper(TM) scissors menacingly
chromatic Do it.
pmichaud muwahahahaha! 20:02
NotFound And fast!
Eevee oh dear. just don't run with those
pmichaud I never "run", I just "execute" with them.
Eevee groan
pmichaud ("with extreme prejudice" :-)
Eevee I concede pun defeat 20:03
NotFound Off with his head!
pmichaud I have Code Ripper(TM) scissors. NotFound appears to have a Code Ripper Chainsaw(TM)
of course, I think I'm the only one who can really do a "Texas Chainsaw Massacre." :-P 20:04
chromatic Wow. You can paste HTML into Blogger and it will still mangle your HTML all to pieces. That's... useful.
particle chromatic: you left in 'aloha' and 'mahalo' in the announcements?
NotFound pmichaud: I think that my LOC contribution will overflow negative integer range.
pmichaud NotFound: I think that negative code contributions should be multiplied by a factor of -2 20:05
DietCoke fixes a C bug. woot.
chromatic I pulled them out of some of the announcements. 20:06
DietCoke chromatic: can you vet my patch on #54474 ?
particle chromatic: the file to fix is tools/util/templates.json
DietCoke ... why is aloha in that file? 20:07
DietCoke glares at particle for leaving an surprise.
NotFound pmichaud: and the bonus added is to close several tickets about fixing the dead functions.
pmichaud NotFound: indeed, bonus
chromatic ahhhhhhhh overload! 20:08
particle ...that's what happens when you release parrot while on vacation...
DietCoke chromatic: ahhhhhhh freakout? 20:09
Eevee we need to upgrade chromatic to quad-core
DietCoke ... that's like having a 64-bit leo. 20:10
NotFound Or clone it and make a cluster.
chromatic Mostly I'm trying to figure out 1) why Blogger helpfully inserts extra <p></p> tags in your HTML when you've already provided it with valid HTML and doing so actually contravenes the HTML specification and 2) who to stab in the face to get it to change.
DietCoke did you use html mode or helpful edit mode? 20:11
particle robrt's probably the closest target
DietCoke you could join google and work on it on your skunkworks.
skunkworks?
bad purl.
purl I'm not bad, I'm just coded that way
chromatic I used HTML mode. It mangled my HTML. "OH, a newline after a paragraph tag! That must mean to add another paragraph tag! Hooray! Oh, a newline in an open tag! What could possibly go wrong if I add a paragraph tag there! Nesting tags within tags is FUN!" 20:12
Eevee s/\\n//g?
chromatic Yeah, because whitespace is significant in HTML. 20:13
Eevee or replace with a space, or add <!-- at the end of a line and --> at the start of the next, or some similar such terrible hack 20:16
or blogger might have some switch or custom tag to disable its own autoformatting, like lj does
chromatic I'm going to go with the face-stabbing, just because it's less of a hack.
Coke: your patch looks fine.
PerlJam chromatic: looks like you'll just have to take time out of parrot/$job/etc. to write a nice blog tool.
chromatic: you could even write it in parrot/perl6 :)
NotFound Are you embedding whitespace code in html?
DietCoke chromatic: danke. 20:17
chromatic Between tags, yes.
NotFound Will be nice: <script lang='whitespace'>... No need to hide the code for old browsers. 20:18
dalek r27683 | chromatic++ | trunk: 20:21
: [tools] Added next release date to Crow's release notes template.
: Added more Haole to the release text.
diff: www.parrotvm.org/svn/parrot/revision?rev=27683
Tene So, if I wanted to contribute posts to planet.parrotcode.org or wherever else someone wanted blog-like contributions, where should I do that? 20:23
DietCoke planet is an aggregator. 20:24
so if you have a blog where you post parroty things, let me know.
creating a blog on blogspot is easy. don't know if allison intended all committers to have access to parrotblog, either. 20:25
allison parrotblog is pretty open
Tene use.perl.org/~tene/journal/ -- I'd post here if it were aggregated there. 20:26
allison I've been inviting anyone who wants to blog there
japhb wishes he had time to *both* contribute code *and* prose. Sadly, I don't really have time for either one, so I'm going with code for now.
DietCoke I'd aggregate there if you posted there. =-)
dalek r27684 | chromatic++ | trunk:
: [PMC] Made lower method return a STRING, not a boxed PMC. (Patrick Michaud, RT
: #54522).
diff: www.parrotvm.org/svn/parrot/revision?rev=27684
allison I keep meaning to post, but don't get to it
DietCoke I did one!
DietCoke sits on his haunches and expects a biscuit. 20:27
allison gives coke a biscuit
PerlJam allison: so how does one get a login to parrotblog?
DietCoke you need a google id, IIRC. 20:28
(it's just bloger)
allison let me know what your gmail account email address is
DietCoke er, blogger.
allison and I'll send out an invite
pmichaud I'd like to post to parrotblog.org, if I can't already.
dalek r27685 | fperrad++ | trunk:
: [Lua]
: - bytecode translation : more refactor & clean up
diff: www.parrotvm.org/svn/parrot/revision?rev=27685
DietCoke leopard? 20:29
purl leopard is a panther or not Panther or OS X 10.5 or at www.apple.com/macosx/leopard/
nopaste "pmichaud" at 76.183.97.54 pasted "flamebait: diff t/steps/auto_pmc-03.t t/steps/auto_pmc-04.t" (23 lines) at nopaste.snit.ch/13028
Tene allison: srweeks@gmail.com
pmichaud my gmail account login address is pmichaud@pobox.com 20:30
or else patrick.michaud@gmail.com
Tene Hm. I should set up a google account for an address I check.
pmichaud (if it requires a gmail.com address)
japhb pmichaud: (flamebait) nuke it 'til it glows! Well, one of them, anyway 20:31
allison tene & patrick: invites sent
pmichaud: it does require a gmail.com address
(it's a single signon, so your gmail account is your blog account)
tene: I just have my gmail account forward to my main email account 20:32
Tene looks like I can sign up for blogger with my non-gmail google account. 20:33
allison: see if you can send an invite to tene@allalone.org
20:33 dcat_ joined
allison tene: apparently yes 20:34
Eevee hm. rakudo.org does not appear to be fed to planetsix.perl.org 20:35
DietCoke tries to figure out what the difference is between SELF.get_integer() and VTABLE_get_integer(INTERP, SELF);
I don't think I have the keys to planetsix, just planetparrot
chromatic DietCoke, I think they're the same; the methody-looking form rewrites to the macro form.
NotFound You came from another planet? 20:36
DietCoke just verified that, yah.
I have no problem with Vaso;u
Vasily's patch on 54474, then. 20:37
allison DietCode: see lib/Parrot/Pmc2c/MethodEmmitter.pm
DietCoke (but having multiple syntaxes is just confusing)
allison Emitter, that is
Eevee subscribes to planetparrot
rss++ 20:38
Tene DietCoke: could parrotblog be aggregated on planet.parrotcode.org?
DietCoke It could, sure. that might almost make sense. =-)
jonathan chromatic: I think the first form maybe does the call on real_self, but only if DYNSELF became SELF already, which I forget it happened yet. 20:39
DietCoke jonathan: it did.
allison DietCoke: it didn't used to be the same, but when DYNSELF became SELF, then only STATICSELF is odd and interesting
DietCoke ah. 20:40
how confusing!
allison At the moment, there's not much need for the macro, but it can be used to hide magical inheritance from HLLs to low-level pmcs 20:41
DietCoke While I'm in here, should I remove chip's blog? leo's blog?
allison are they posting parrot stuff? or anything?
jonathan allison: I think at the moment, we're not fully using the new PDD17 model for inheritance, attributes and so on?
allison jonathan: correct, it's partially there, but not fully
jonathan We still instantiate the parent PMC and delegate at the mo... 20:42
OK.
allison for one thing, we haven't eliminated real_self yet
jonathan Right.
Is this on your queue, or is anyone free to take a look?
allison but, we are now generating the low_level attributes, so we're 5 steps closer than the last time we talked about it
jonathan After the current batch of conferences, my schedule is a bit lighter.
allison my queue is non-blocking :) 20:43
take a look, ask questions, do it if you get to it
(though, bytecode is probably a higher priority)
DietCoke parrotblog added to planetparrot. give it a few hours. 20:44
Tene NO I NEED IT NOW
Maybe.
20:45 cjfields joined
spinclad (re flamebait): <flame> repetition is the soul of wit. long established principle. </> 20:47
jonathan allison: Infinoid++ is taking a look at bytecode, though I may well join in helping out there too. :-)
PerlJam jonathan++ 20:48
pmichaud spinclad: yes, I think I heard that somewhere.
allison I've seen that, Infinoid++,
20:58 ruoso joined
dalek r27686 | pmichaud++ | trunk: 21:02
: [core]:
: * Remove the 'Super' PMC, resolving RT#53968.
diff: www.parrotvm.org/svn/parrot/revision?rev=27686
21:04 davidfetter joined
DietCoke those auto_pmc tests should probably not have a list of all the pmcs in them. 21:07
21:19 peepsalot joined 21:38 teknomunk joined
chromatic How far away is fail() support in Rakudo? 21:46
cognominal thie parentheses are are scary cuz they are used in p5 style :( 21:48
> my $x = '0'; if $x ne $x { say 'not ok 1' } else { say 'ok 1' }
ok 1
> my $x = '0'; if ($x ne $x) { say 'not ok 1' } else { say 'ok 1' }
not ok 1
>
NotFound Will be acceptable to rename const_cast to PARROT_const_cast? 21:54
jonathan chromatic: Do you know where fail is actually defined? Just went to check in the synopses to see exactly what it requires, and didn't find a definition...
S29 just says, "TODO: Research the exception handling system." 21:56
Ah, S04 seems to define it under exceptions 21:58
pmichaud doesn't it just throw an exception?
moritz not necessary
jonathan Not really, that's die
It may under use fatal 21:59
moritz onlw when the "fatal" pragma is active
pmichaud ah.
moritz if not, it returns an "interesting undef" value
jonathan Otherwise, it "registers" itself with $!
moritz which is like an unthrown exception
jonathan And returns undef
pmichaud cognominal: that's definitely a rakudo bug -- feel free to submit it as a bug report.
cognominal that's always the same bug in another guise 22:00
pmichaud $x ne $x and ($x ne $x) should return the same.
jonathan Indeed.
I'll try and look at that tomorrw...that's nasty.
pmichaud I'm looking now. 22:01
jonathan Oh, cool.
I would, but my slides are still not sufficiently written.
pmichaud pmichaud@orange:~/parrot/trunk/languages/perl6$ ./parrot perl6.pbc
my $x = '0'; if $x ne $x { say 'not ok 1' } else { say 'ok 1' }
ok 1
my $x = '0'; if ($x ne $x) { say 'not ok 1' } else { say 'ok 1' }
ok 1
so I'm not seeing the same bug.
cognominal: perhaps you still have a local () patch applied?
jonathan cognominal: You haven't still got some local diffs, e.g. theunapplied....
pm beat me! 22:02
cognominal I check
chromatic NotFound, where are you looking at renaming the macro?
NotFound chromatic: include/parrot/parrot.h 22:03
22:04 cotto_work joined
NotFound I was hoping that this change will solve a problem I have compiling with g++, but It doesn't 22:05
chromatic Hm, that *is* pretty global. I was just going to ask about C++ problems with it.
dalek r27687 | pmichaud++ | trunk:
: [pge]:
: * <?foo> is now a zero-width match instead of a non-capturing subrule.
: * This brings us back into closer compliance with S05.
: * Resolves RT#53834.
diff: www.parrotvm.org/svn/parrot/revision?rev=27687
cognominal yes, that my stupid patch that is kiling me
NotFound Potantially any c++ program embedding parrot will have problems with that macro. 22:06
By the way, there is a 'new' identifier in dod.h 22:07
dalek r27688 | pmichaud++ | trunk:
: * Forgot to update DEPRECATED in previous commit.
diff: www.parrotvm.org/svn/parrot/revision?rev=27688
chromatic I don't see it. 22:08
Oh, in some of the generational APIs.
NotFound In parrot_gc_gms_wb 22:09
And parrot_gc_gms_wb_key 22:10
22:11 DurableAlloy joined
chromatic I'll fix those up in a minute; I'm editing POD right now. 22:13
NotFound char ** const strings; 22:15
g++ don't like this. 22:16
dalek r27689 | chromatic++ | trunk: 22:19
: [docs] Minor edits to PCT Getting Started guide, prompted by the desire to fix
: its indexing on the CPAN.
diff: www.parrotvm.org/svn/parrot/revision?rev=27689
r27690 | chromatic++ | trunk: 22:21
: [GC] Changed C<new> parameter to C<_new> in a couple of GMS functions to help
: C++ compilation. Reported by NotFound.
diff: www.parrotvm.org/svn/parrot/revision?rev=27690
22:21 IllvilJa joined
NotFound I have a strange problem compiling with g++ 4.2.3, several functions claims to have a nothrow () declaration that does not match his definition. 22:25
All of them have one parameter, attributed as nonnull. 22:26
22:26 ruoso joined
dalek r27691 | pmichaud++ | rakoo: 22:26
: Create a branch for refactoring rakudo base classes and objects.
diff: www.parrotvm.org/svn/parrot/revision?rev=27691
moritz pmichaud++ # "rakoo" ;-) 22:27
pmichaud just registered rakoodo.com :-) 22:31
moritz ;)
22:35 rdice joined
Juerd What's rakoodo, besides confusing? 22:47
pmichaud I'm just reserving the name is all.
"rakoo" is simply my branch where I'm refactoring the oo components of rakudo. 22:48
Juerd Ah :)
pmichaud it'll disappear in a day or so.
Juerd Expensive to register a domain for just a day or so :D
pmichaud oh, I might use it in the future for something. :-)
Juerd It's interesting to see that you pick .com, whereas rakudo.org is not .com
pmichaud oh, I grabbed .org as well. :-) 22:49
Juerd Heh
You're thorough :)
pmichaud I've had experience with not grabbing domain names when they were available and regretting it later.
22:49 DurableAlloy left
NotFound Is there some way to disable jit capability? 22:49
pmichaud probably an options to configure.pl 22:50
try Configure.pl --help
NotFound There is --jitcapable to enable it.
Juerd pmichaud: I have that same experience. 22:51
That's why I got juerd.com/net/org even though I don't currently use them.
Well, they're active but I don't use them in communication
jonathan pmichaud: It appears there will be hackathon after YAPC::EU over in Sweden, arrive Sunday, hack for few days (up to Wed) or whatever of it works for you, if you're interested. 22:53
Just got email about it.
I'd forward, but your mail server always rejects my emails. :-|
pmichaud jonathan: excellent, thanks. I'll see if I can fold it into my travel plans.
which server rejects them? pobox.com?
or gmail?
purl gmail is hotmail with less suck. NOTE: It is NOT repeat NOT (repeat: not repeat not) mutt with less suck. It's webmail, folks. Brilliant, well-executed webmail, but still webmail. or Google-powered suck or the best webmail Masque has ever used. n.b., this is like saying, "It's the best wood I've ever eaten!" or scary or cartmanmail or spawn of evil
Juerd gmail is also still webmail :) 22:55
purl okay, Juerd.
jonathan pmichaud: pobox
pmichaud hrm.
Juerd I had really expected gmail to get its own mail client. 22:56
Infinoid gmail has imap support
Juerd Infinoid: Oh, that's interesting.
Eevee Juerd: it has one, it's just written in html/js
Infinoid mail.google.com/support/bin/answer....swer=75725
pmichaud jonathan: you could send to patrick.michaud@gmail.com, or I could whitelist you directly into pmichaud@pmichaud.com
Juerd Eevee: Where can I download it? :) (I meant offline client)
pmichaud (all of those addresses go to the same place.) 22:57
Juerd Infinoid: How do they push the advertisement through imap then?
Eevee Juerd: well, then you're in trouble :P but imap should be fine if that's what you want
although
I can see prism + gmail + google gears working in interesting ways
Infinoid Juerd: apparently, they don't. :) 22:59
Juerd Infinoid: Hm...
pmichaud afk # swimming 23:01
23:11 tetragon joined
confound jonathan: why does pobox reject your email? 23:12
jonathan confound: Because my (in data center) mail server has numbers in it's hostname, so it thinks it's somebody's home machine that's been 0wned. 23:16
And I ain't got onto getting reverse DNS entry changed yet.
NotFound Finally, with some tricks and using --without-gmp I've been able to compile with g++ :) 23:17
confound jonathan: what's your mail server's hostname? 23:18
NotFound ...but fail a lot of tests. 23:22
chromatic NotFound: fail as in crash, or just the wrong answer? 23:24
NotFound Failed to compile. 23:25
23:52 davidfetter joined 23:54 bacek_ joined
bacek_ morning. 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!