Or bribe^Wcontact Andreas directly
Not sure. I don't really like either.
Nicholas Clark
a behaviour of the hidden invoke (of the exception handler) that
> I am unaware of?
This got Warnocked, didn't it?
Nicholas Clark
or may not reflect the current code.
This got Warnocked. I think it would be a good idea to put abstracts
into the documentation, a\as currently it does dive straight into details.
Nicholas Clark
On Sat, Sep 20, 2003 at 08:06:12PM +0100, Steve Purkis wrote:
> On Saturday, September 20, 2003, at 06:11 pm, Steve Fink wrote:
>
> > On Sep-20, Nicholas Clark wrote:
> >> On Sat, Sep 20, 2003 at 10:02:34AM -0700, Steve Fink wrote:
> >>
> >>> Can&#
write all the way up the stack from the kernel read to the
PMCs presenting data at language level, so that for read only grepping
of files there is no data copying at all. (And hopefully the file at the
bottom is memory mapped)
Nicholas Clark
eq S0, S1, equal
> print "not "
> equal:
> print "equal\n"
>
> index I0, S0, S1
> print I0
> print "\n"
>
> end
>
> yields the output:
> --------
> equal
> -1
>
> which doesn't seem quite right.
Looks like a bug to me, but I'm no expert on where to start searching for
it.
Nicholas Clark
omething that
happens after file handle construction (hence there can be a stack by open
time)?
Nicholas Clark
inated string can do the strlen
themselves. (Or we could provide a helpful macro)
Nicholas Clark
bious
Test returned status 1 (wstat 256, 0x100)
DIED. FAILED test 4
Failed 1/4 tests, 75.00% okay
Did MANIFEST get committed without some files?
Nicholas Clark
On Sun, Sep 21, 2003 at 05:49:41PM +0100, Nicholas Clark wrote:
> I'm seeing this failure on a clean checkout:
>
> t/src/manifest.NOK 4# Failed test (t/src/manifest.t at line 38)
> # Missing files in Manifest:
> # languages/jako/examples/python.jako
> #
couple of places it was using the
>typedef of Parrot_Interp.
> - pdump.c contained declarations for functions in packdump.c. I have
> moved these to packfile.h.
Thanks, Applied
Nicholas Clark
n-a-stick :-)
> Argh.
Thanks for all your toils in getting a release out.
Nicholas Clark
nformation about which quadrant the angle is in
(to save you working it out)
Er, but this doesn't really add to the informed debate.
[except that maybe Dan should be suggesting atan2() rather than arctangent]
Nicholas Clark
nderstanding is that the relay from the mailing list to google
groups is one way - ie messages sent to google groups never get back
here.
Is there a good way to make this clear on google groups?
Is google able to flag groups as "read only"?
Nicholas Clark
ldn't guess)
"Yet another certifiable Austral^H^Hian" :-)
Nicholas Clark
tml?id=24029 >
>
>
> call_list.txt has two "i 4i" lines.
Mmm yes. Without this parrot wouldn't build for me on PPC Linux
Thanks applied
Nicholas Clark
Either that or put them in a subdirectory, and rig the include path
to have that subdirectory in it.
We're making the top level project source directory the "subdirectory"
at work (ie point -I one higher than you might first think) and it seems
to "namespace" our includes quite nicely.
Nicholas Clark
h every class.
However, would it be possible to schedule the other classes for lazy loading
(à la perl5's AUTOLOAD), so that the functions behave as if they were
defined at the correct time, but actually pull in the needed bulky libraries
only if invoked?
Nicholas Clark
er for a serialisation call while things are only partially
loaded?
I've got no real idea on any of these, apart from a hunch that assigning
numbers to classes that aren't loaded (yet) is going to be needed for
other things.
Nicholas Clark
re violates one of the constraints:
> On Monday, September 29, 2003, at 01:36 PM, Robert Spier wrote:
> > Restrictions:
> > The script cannot access any other files/filesystems/directory
> > listings besides the data in the files provided on the command
> > line.
Nicholas Clark
n and multiplication, never from addition or *in this case)
subtraction.
Nicholas Clark
On Thu, Oct 02, 2003 at 11:52:48AM -0400, Dan Sugalski wrote:
> On Thu, 2 Oct 2003, Nicholas Clark wrote:
>
> > On Thu, Oct 02, 2003 at 10:35:29AM +0200, Leopold Toetsch wrote:
> > > Adam Thomason <[EMAIL PROTECTED]> wrote:
> >
> > > > The one remai
nstructions, but they're all
inside blocks that are only true on dosish platforms.
Nicholas Clark
e, thanks
Yes, but if you have to go via another PMC register then you may spill.
Whereas it seemed that the type retrieval ops went direct to S and I
registers, which may save spillage. (at a cost of more ops)
Nicholas Clark
1: pbc.h: No such file or directory
imcc.y:22: parser.h: No such file or directory
This is with gcc version 2.95.4 20020320 [FreeBSD]
Have the rules about where to find header files in directories relative
to source files been changed between gcc 2.95 and gcc 3?
Nicholas Clark
OCTL_CMDGETBUFSIZE:
> if(b) return b->size;
> else return -6;
>
What are all these magic numbers about?
Nicholas Clark
On Fri, Oct 10, 2003 at 12:35:23PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > I'm having trouble testing this because I can't get CVS parrot to build on
> > my (friend's) freebsd box. I think that this ought to restore
racters, the
> list goes on.
Sorry. Wasn't clear. The PIOCTL macros are (to me) the correct way of doing
something.
-6 and -3 are magic numbers. They have no context as to what they might be.
> At 04:46 PM 10/10/2003 +0100, Nicholas Clark wrote:
> >On Fri, Oct 10, 2003 at 02:43:1
FreeBSD make doesn't understand those % pattern rules
2: gcc 2.95 resets its idea of the source file directory on a #line
directive, and as a consequence can't find other files to include
in that directory"
Not sure how to solve this short of changing directory.
Nicholas Clark
installing the opcode dispatch table at runtime,
> but we must reserve the opcode number range, and the assembler must
> at least be able to query the dynamic lib for valid opcode names. When
> the first unknown opcode is to be run, then the real functions could be
> put in place.
Or am I getting confused and all this lookup is done by name?
Nicholas Clark
nning on CG or some other
low startup cost core until then)
or
b: signal a (partial) re-JIT of something/everything after a library/
everything has loaded?
I've no idea how easy this might be. Or sensible.
Nicholas Clark
be a throw away question. As I understand it many hostname
functions were changed or tweaked for IPv6, and coping with it now is
better than retrofitting it later)
Nicholas Clark
On Fri, Oct 31, 2003 at 12:09:14AM +, Nicholas Clark wrote:
> On Thu, Oct 30, 2003 at 01:54:24AM -0500, Melvin Smith wrote:
> > Parrot fetched its first web page tonight. Its a baby step, but hey... :)
I forgot to say:
Hey, cool! Nice work.
Today web pages, tomorrow self propagat
, and set the reply to (to ensure responses move there)
I'm not sure of the best answer to your question, but someone on
perl5-porters should be able to answer it.
Nicholas Clark
the same. I presume that there
are some encodings where two different binary representations are considered
"equal", hence we can't blindly assume that a byte compare is sufficient.
Nicholas Clark
27;ve no idea if using an iterator breaks some or other design rule.
Nicholas Clark
FFF (Unicode as code-points up to 0x10FFF), as either two 16 bit
> surrogate code points encoded as two 3 byte UTF-8 code sequences or as
> a single value encoded as a single 4 or 5 byte UTF-8 code sequence.
Is it legal to encode surrogate pairs as UTF8? Or does that count as
malformed UTF8?
Nicholas Clark
acros,
such as not always giving as good optimisations as macro-re-written code.
Is this something worth taking further?
(Ulterior motive - I'm curious whether doing this would also help perl 5)
Nicholas Clark
if that fails retrying with all the found
headers (in order). The second try means that netinet/in.h is now "found"
on FreeBSD, and I can build on FreeBSD 4.8
I believe that this resolves the bug
Nicholas Clark
t;
>
>
> This attached patch adds support for OpenBSD 3.4 (which uses ELF), without affecting
> older versions using a.out (hopefully). This cures the tinder build failure on
> phalanx. Thanks to Daniel for help tracking down the necessary symbols.
Thanks, applied.
Nicholas Clark
de to better alignments based
on the amount of padding that would be needed. These defaults mean that
changing the size of earlier parts of the object file can affect the
alignment (and hence speed) of loops you didn't change. This is very
confusing.
Nicholas Clark
its':
src/dod.c:755: `cur_arena' undeclared (first use in this function)
src/dod.c:755: (Each undeclared identifier is reported only once
src/dod.c:755: for each function it appears in.)
src/dod.c
*** Error code 1
The appended patch cures it (and all tests pass) but I'm not sure if i
On Sat, Jan 10, 2004 at 08:39:51PM +0100, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > src/dod.c: In function `clear_live_bits':
> > src/dod.c:755: `cur_arena' undeclared (first use in this function)
>
> > The appended patc
esting.)
My tinderbox running on FreeBSD definitely didn't have it, and the OpenBSD
tinderbox returned from burning to success at the same time as mine.
(Not that I've checked the actual case of its failure, but I think only
one patch got committed around then)
Nicholas Clark
ness has just one
> > result summary.
>
> Fine, thanks - applied.
It broke my tinderbox, because it assumed that . was in $PATH
I committed a fix.
Nicholas Clark
> configure.
There was a time when the first perl in my $PATH was /bin/false
I might make that the case on my tinderbox.
Muahahahahaha.
Nicholas Clark
On Sat, Jan 10, 2004 at 10:30:47PM +0100, Stéphane Payrard wrote:
> retry
You're trying to attache a file with a name ending in .t ?
They get eaten. I forget why. And I forget why it's not been possible
to change the configuration on the list software to tell it that it's
on
On Sun, Jan 11, 2004 at 07:33:12PM +0100, Stéphane Payrard wrote:
> Also the following patch exercises all the variations including
> the susmentioned access thru a key pmc that may contain either a
> string or an int.
Thanks, applied
Nicholas Clark
> <>
Have the LZW patents(*) expired everywhere yet?
Nicholas Clark
* Remember that those smart, diligent folks at the US Patent Office granted
both the Sperry Corporation and IBM a patent on the same algorithm at about
the same time.
pecifically so that pipe() could be implemented on it and hence be select()
able.
Given that I wrote it, I don't see any copyright leakage problem here :-)
Nicholas Clark
On Sun, Feb 29, 2004 at 05:36:22PM +0100, Leopold Toetsch wrote:
> Currently working on 10a of RELEASE_INSTRUCTIONS everything should be
> done, so let the checkins begin.
Woohoo!
Well done.
I see it's all over use.perl now too.
Nicholas Clark
different places on OS X. Not tried other
systems yet. Still fighting AIX]
We're not sure how to track this one down further - at the moment the best
plan seems to be to apply the above patch to ponie, and make a snapshot
release with it in.
I've attached the diff between TEST run for th
On Thu, Mar 11, 2004 at 10:33:24PM +0100, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > If parrot's garbage collector is changed from the default (compacting, IIRC)
> > to the either libc or malloc, then ponie only fails 6 tests.
>
On AIX, what's the difference between cc_r and xlc_r?
And why does parrot's hints file go for xlc_r, whereas perl5's goes for cc_r?
This is causing pain for ponie. Is there any reason not to pick the same one
for both?
[yes, 3-way cross post, but I think it's justified]
Nicholas Clark
On Tue, Mar 16, 2004 at 10:23:34PM +0200, Jarkko Hietaniemi wrote:
> Nicholas Clark wrote:
>
> > On AIX, what's the difference between cc_r and xlc_r?
>
> See /etc/xlc.cfg.
>
> I vaguely remember that's it's the cc_r that's guaranteed (well, *mor
on after a cvs up on the parrot
directory, but this may not be.
Nicholas Clark
inadvertently ignored, but it is something we (the perl comminity)
aren't getting right yet, and it is something we should fix.
As to the original bug you report - I no longer have access to any SuSE
machines, but I did build perl from source on a SuSE 8.1 box within the
past 3 months and didn't encounter the problems you encountered. Strange.
Hopefully someone else has access to SuSE and can replicate your problems
Nicholas Clark
ble to do
"whatever" and return a file handle that might have 1 or more
transformation layers on it, then I think it covers most possibilities.
Nicholas Clark
v6. Seems to work from here. Only this is the only IPv6 host
that I have a shell account on.
Nicholas Clark
parrot build tree.
Which can cause test failures.
(IIRC some of MakeMaker's tests can do this, and I've yet to supply schwern
with a patch)
Nicholas Clark
reter,
rather than the full path).
But all this is from memory, and in turn for #! invocation one can always
parse the #! line to work out where the interpreter was (mmm. race
condition)
Nicholas Clark
`typedef struct Parrot_Interp*Parrot_Interp'
Pain being due to these two:
struct Parrot_Interp;
typedef struct Parrot_Interp *Parrot_Interp;
This doesn't seem right.
Nicholas Clark
On Sat, May 08, 2004 at 11:40:10AM +0100, Nicholas Clark wrote:
> So I've fixed that and now it all runs in reasonable time. (As for all its
> inherent horrors, perl's reference counting pretty much gets it right as to
> when it's time to free up something)
Reasonabl
I should check my aliases before hitting send, shouldn't I?
- Forwarded message from Nicholas Clark <[EMAIL PROTECTED]> -
Mailing-List: contact [EMAIL PROTECTED]; run by ezmlm
Precedence: bulk
From: Nicholas Clark <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED], [EMAIL PROTE
On Fri, May 07, 2004 at 11:56:10AM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > 0x0002e354 in pobject_lives (interpreter=0x1000200, obj=0xd13a68)
>
>
> Ok, that'
r
as I can tell nowhere in any documentation files, and there aren't many
comments about it in include/parrot/pobj.h
I don't find any documentation saying "if you do this you'll need to do that
in a destructor else you'll leak".
Am I missing something obvious?
Nicholas Clark
7 0.00 0.00 Perl_macro_SvPVX
0.03 1270.87 0.38 1632298 0.00 0.00 Perl_pp_regcomp
0.03 1271.23 0.37 1808077 0.00 0.00 Perl_pp_match
Any suggestions on what stats to gather to try to work out where things are
going wrong?
Nicholas Clark
;t seem to fail on Linux (with notgcc 2.96) so I'm somewhat at a loss
as to what to fix.
Nicholas Clark
On Fri, May 07, 2004 at 06:57:24PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > With all the perl scalars now indirecting through PMCs, Ponie's performance
> > has dropped. Not surprising, in the general case, but the performance of
>
On Fri, May 07, 2004 at 08:17:05PM +0200, Leopold Toetsch wrote:
> Ah yes. Needs
>
> #if ARENA_DOD_FLAGS
>
> around. Fixed
Thanks. Everything compiles again now, and all tests pass
(after I do the little .so -> .dylib dance)
Nicholas Clark
ho is in a position to
structure their own code/project to set the stack top once and in the
correct place)
How should this be?
Nicholas Clark
diff -d -u -r1.26 extend.c
--- src/extend.c3 May 2004 12:29:02 - 1.26
+++ src/extend.c7 May 2004 20:15:20 -
@@ -789,6 +789
On Fri, May 07, 2004 at 03:11:54PM -0600, Luke Palmer wrote:
> Nicholas Clark writes:
> > Thanks. Everything compiles again now, and all tests pass
> > (after I do the little .so -> .dylib dance)
>
> I'm new to OS X. Might you describe said dance?
We haven
On Fri, May 07, 2004 at 10:09:10PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > The PMC only used the PMC_data member to store a pointer on to a perl5
> > SV body.
>
> That's suboptimal :) Please have a look at the timings I
On Sat, May 08, 2004 at 10:01:56AM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> [ half second DODs ]
>
> > 0.07 1266.73 0.86 9459863 0.00 0.13 get_free_object_df
>
> > 0.05 1268.18 0.62 1240633
On Fri, Apr 30, 2004 at 01:54:58PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > 1: Right now, would it be possible for parrot only to install its signal
> >handlers when it starts the runloop?
> >(given that ponie isn'
t you link
against existing (unchanged) C code.
It's possibly not an issue long term for ponie (as a stand alone interpreter),
except that it would mean that ponie couldn't be embedded in the way that
perl5 currently can be embedded.
Nicholas Clark
On Fri, Apr 30, 2004 at 06:25:50PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
> > This is true. And yes you need to set a stack top if you're expecting the
> > stack walking to find things you own. But I'm thinking about the c
l when parrot is built as ponie
(which as far as I can tell involves 3 more PMC classes being added to the
parrot source tree)
And where should I start to dig to work out what is wrong? I tried
perl -I lib t/pmc/dumper.t
from the parrot directory, and earlier tests (6 and 7) fail but test 9 passes.
Nicholas Clark
to save us a couple kilobytes.
Wibble.
I hacked perl's Configure instead. It felt like the less painful option.
Nicholas Clark
l foul of the above.
Effectively the embedder is playing by the same rules as Leo said that we
have to have for a nested call to the runloop - anything calling the runloop
must ensure that all PMCs are anchored.
Nicholas Clark
On Mon, May 03, 2004 at 04:36:38PM -0400, Dan Sugalski wrote:
> At 9:33 PM +0100 5/3/04, Nicholas Clark wrote:
> >On Mon, May 03, 2004 at 10:46:28AM -0400, Dan Sugalski wrote:
> >
> >> 3) The embedding wrapper is responsible for setting and resetting the
> >> t
e may have to throw in a config/auto/linker.pl file to
> identify the linker, along with linker-specific config modules.
Which implies that platforms may exist where it's not possible to do this.
(surely?)
Nicholas Clark
r-e-mail session with someone at Boeing,
but I don't think that we managed to get all the bugs nailed down.
Nicholas Clark
/embed.c:742
#11 0xf960 in Parrot_runcode (interpreter=0x1000200, argc=1, argv=0xb784) at
src/embed.c:676
#12 0x00004284 in main (argc=1, argv=0xb784) at imcc/main.c:556
Nicholas Clark
On Thu, May 06, 2004 at 04:43:53PM +0200, Leopold Toetsch wrote:
> Nicholas Clark <[EMAIL PROTECTED]> wrote:
>
> > +$(LIBNCI_SO): $(SRC)/nci_test$(O)
> > $(LD) $(LD_SHARED) $(LDFLAGS) \
> > $(LD_OUT)$@ $(SRC)/nci_test$(O)
>
> Win32 additionally
On Thu, May 06, 2004 at 06:21:23PM +0200, Leopold Toetsch wrote:
> Nicholas Clark wrote:
> > 0 loadlib P1, "libnci" - P1=NULL,
> > 3 dlfunc P0, P1, "nci_pi", "pi"- P0=NULL,
> > P1=ParrotLibrary=PMC(0x200f7f0), , DOD
> &
On Mon, May 17, 2004 at 05:21:05AM -0700, jquelin @ mwinf0401. wanadoo. fr wrote:
> dflguqfgulqf g
So the new parrotbug works? :-)
> LANGUAGE=en_US:en
?
Isn't that an oxymoron?
d hence whether this is
still the case.
It doesn't look like anyone has added anything to the JIT recently:
$ ls -l jit/arm/
total 72
drwxr-xr-x 5 nick nick170 6 May 17:42 CVS
-rw-r--r-- 8 nick nick 27256 19 Oct 2002 core.jit
-rw-r--r-- 8 nick nick 3561 23 Apr 10:20 exec_dep.h
-rw-r--r-- 8 nick nick 37553 23 Apr 10:20 jit_emit.h
Nicholas Clark
o the code point splitting) and writing the values out as
32 bit quantities there and then will take virtually no more CPU, but
save lots later.
Or is this now all gone, because there will be Unicode everywhere and
strings will get converted at input IO time?
Nicholas Clark
ad pitching exceptions at conversion time?
Nicholas Clark
ly for fixed width, UCS 32 (IIRC).
The only thing that might be useful to cache on a UTF8 string is the highest
code point seen, so that we know whether to unpack to 8, 16 or 32 bit without
a scan. Presumably we can find this when we input validate on the
"conversion" from binary to UTF8.
Nicholas Clark
terested in is finding the closest, not the distance
itself, what's wrong with using square of geometric distance?
(Given that the squares of positive numbers sort in the same order as the
numbers themselves.)
One can also do weightings with this still in integer arithmetic, if
weightings are small positive integers. (And we don't overflow)
Nicholas Clark
be malloc()ed, rather than grabbed from something
parrot can pass onwards through PMCs? Is the intent to keep the IO system
isolated from most of parrot, so that it still works during VM
setup/teardown?
Am I guilty of premature optimisation here?
Nicholas Clark
arrot_getenv("ICU_DATA_DIR", &free_data_dir);
Is ICU_DATA_DIR something the ICU folks define? Or something we define?
And if the latter, shouldn't it be called PARROT_ICU_DATA_DIR?
Nicholas Clark
In the meantime, however, this patch summarizes the
> current state of affairs.
The work around for (2) could be to bundle File::Spec 0.87 from CPAN.
For that matter the work around for (1) could be to bundle Math::BigInt
from CPAN. :-)
Nicholas Clark
On Wed, Jun 09, 2004 at 09:46:52AM -0700, chromatic wrote:
> On Wed, 2004-06-09 at 03:53, Nicholas Clark wrote:
>
> > The work around for (2) could be to bundle File::Spec 0.87 from CPAN.
> > For that matter the work around for (1) could be to bundle Math::BigInt
> >
d by swearing and cursing and time wasted grepping to find out what
a STRING is.
Is there any reason not to have a coding standard that typedef should be
used whenever possible (and non-uses must have their reason commented?)
Nicholas Clark
On Fri, Jun 11, 2004 at 06:23:05AM -0700, Nicholas Clark wrote:
> buffer. I'm not sure if the check on meeting the delimiter midway is now
> unnecessary. It's not clear if that code is only there to deal with
> terminating when the end delimiter is encountered, and the end
On Sat, Jun 12, 2004 at 02:57:18PM -0700, Nicholas Clark wrote:
> I suspect a memory corruption bug, as it it doesn't look like the specific
> arguments to this malloc are bogus. I can't recreate this problem at all
> on x86 FreeBSD or on OS X. I find the x86 FreeBSD
On Sat, Jun 12, 2004 at 11:32:16PM +0100, Nicholas Clark wrote:
> Er, oops. OS X wasn't uptodate. Test 8 crashes. Will the guard malloc I get
> further:
>
> (gdb) print data[512]
> Cannot access memory at address 0xb30a3000
> (gdb) wh
501 - 600 of 903 matches
Mail list logo
