Hey all.
I'm working on implementing the ord(i,s) and chr(s,i) opcodes I talked about
earlier, and I noticed what I consider a bug: there exist no transcode
functions to or from native.
Also, the diagonals (identy transforms) don't exist. This means that you
have to explicitly check that you
Bryan --
> On Friday 12 October 2001 12:01 am, Bryan C. Warnock wrote:
> > The attached patch...
>
> {Sigh}. Withdrawn.
Is this no longer applicable? Or have things diverged too much since you
made the patch? I went back and looked at it, and it seemed reasonable
enough, although I don't know
On Friday 12 October 2001 12:01 am, Bryan C. Warnock wrote:
> The attached patch...
{Sigh}. Withdrawn.
--
Bryan C. Warnock
[EMAIL PROTECTED]
perl Configure.pl breaks on my win98. The offending line(s) are :)
> # Temporary hack
> system("make include/parrot/vtable.h");
(I just got it out of cvs, but if I am doing sth stupid please tell me)
So I decided to try to fix it.
My changes are
- Configure.pl:
I added a prompt for a make-p
On Wed, Oct 17, 2001 at 12:40:36AM +0100, Tom Hughes wrote:
> sub(i, i, ic)
> sub(i, ic, i)
Should be OK, because I don't think we'd ever distinguish these
two.
--
A language that doesn't have everything is actually easier to program
in than some that do.
-- Dennis M. Ritchi
In message <[EMAIL PROTECTED]>
Dan Sugalski <[EMAIL PROTECTED]> wrote:
> At 05:21 PM 10/16/2001 +0100, Tom Hughes wrote:
> >I have a plan to semi-automate it which I nearly implemted the other
> >day but didn't get around to. Basically the idea is to extend things
> >so an ops file can
On Tue, Oct 16, 2001 at 04:58:47PM -0400, James Mastros wrote:
> There's /got/ to be a better way to write these. I propose making opcodes
> specificly to do these. This is an inner-loop kind of thing.
There will be, of course; the substr solution doesn't really scale to
Unicode. :)
--
We all
On Tue, 16 Oct 2001, Dan Sugalski wrote:
> Ummm. Well. I think that you'll find this pretty much the case already (or
> planned, at least, as some of the bits aren't done yet) so I'm not sure
> that it'd really buy anything other than slower C code.
Wooho!
> Which isn't to say we can't expose the
At 04:48 PM 10/16/2001 -0400, James Mastros wrote:
>What I'm thinking is that if you can compile C code to target the PVM, we
>can have it be /almost/ trivial to run C code with the portablity of
>Parrot, and make C code callable from Perl and vice-versa with only a
>recompile. (Or you can not re
On Tue, 2001-10-16 at 15:58, James Mastros wrote:
> On Tue, 16 Oct 2001, Brian Wheeler wrote:
> > _chr - create a string (S0) with the ascii value of I0
> > _ord - return (in I0) the ascii value of the first character in S0
> There's /got/ to be a better way to write these. I propose making opcod
At 03:58 PM 10/16/2001 -0500, David M. Lloyd wrote:
>On Tue, 16 Oct 2001, Dan Sugalski wrote:
>
> > PMCs that are shared between interpreters are responsible for
> > maintaining their own internal integrity. To this end, every vtable
> > has a shadow "shared" vtable. When a PMC is shared, a mutex
On Tue, 16 Oct 2001, Brian Wheeler wrote:
> _chr - create a string (S0) with the ascii value of I0
> _ord - return (in I0) the ascii value of the first character in S0
There's /got/ to be a better way to write these. I propose making opcodes
specificly to do these. This is an inner-loop kind of
On Tue, 2001-10-16 at 15:02, James Mastros wrote:
> On Tue, 16 Oct 2001, Dan Sugalski wrote:
> > That's one way to do it, sure. You can always look at a string as a bounded
> > byte buffer. One of the core 'string' types is "series of 8-bit bytes". We
> > couldn't manage JPEG images too well witho
On Tue, 16 Oct 2001, Dan Sugalski wrote:
> Parrot's going with the "one thread per interpreter" model of
> threading.
Yay!
> PMCs will be shareable, but only when explicitly marked as shared. Passing
> an unshared PMC to another interpreter's a fatal error. PMCs *can* be
> cloned when an interp
On Tue, 16 Oct 2001, Dan Sugalski wrote:
> You access the bytes individually the same way you do now. If a string is
> of type "8-bit byte" then a character is a byte, and vice versa.
> >I'm thinking of porting GCC, of course . However, I'm thinking that
> >pretty much any c-like language is go
At 12:35 PM 10/16/2001 +0200, Bart Lateur wrote:
>On Mon, 15 Oct 2001 22:12:58 -0400 (EDT), Dan Sugalski wrote:
>
> >>doing:
> >> save S0
> >> restore S1
> >>
> >>(since there's no set S1,S0)
> >>
> >>binds the registers together, so a change to one is a change to
> >>both...which doesn't hap
Okay, I think a quick sketch of the current plan's in order, since I'm
about to start implementing it. :)
Parrot's going with the "one thread per interpreter" model of threading.
That makes a lot of things safer, since we reduce the amount of stuff
shared between interpreters, and don't try ru
At 04:02 PM 10/16/2001 -0400, James Mastros wrote:
>On Tue, 16 Oct 2001, Dan Sugalski wrote:
> > That's one way to do it, sure. You can always look at a string as a bounded
> > byte buffer. One of the core 'string' types is "series of 8-bit bytes". We
> > couldn't manage JPEG images too well witho
I've written a library of sorts which contains useful routines such as:
_absi - absolute value of I0
_absn - absolute value of I0
_chomp - chomp a string (S0) with a trailing newline
_chr - create a string (S0) with the ascii value of I0
_exit - terminate with a return code of I0
_hex - return i
On Tue, 16 Oct 2001, Dan Sugalski wrote:
> That's one way to do it, sure. You can always look at a string as a bounded
> byte buffer. One of the core 'string' types is "series of 8-bit bytes". We
> couldn't manage JPEG images too well without that. ;)
Hm. How do you convert the bytes into integer
James and Melvin --
> This file is supposted to be autogenerated from core.ops, however the
> dependincies in the makefile are wrong.
I just fixed the dependancy. Hopefully, this problem goes away now.
Regards,
-- Gregor
_
All --
> >AUTO_OP add(i, i, i|ic) {
> > $1 = $2 + $3;
> >}
> >
> >and the opcode reading module would expand the i|ic to create two
> >separate versions of the op. Obviously if two arguments had variants
> >you would get four versions and so on.
> >
> >If people think that's a good solution to
At 02:56 PM 10/16/2001 -0400, James Mastros wrote:
>On Tue, 16 Oct 2001, Dan Sugalski wrote:
> > We sort of do, if you treat string contents as a buffer of bytes rather
> > than characters. Raw memory sort of lives a step below the interpreter at
> > the moment, though I can see uses for that not
On Tue, 16 Oct 2001, Melvin Smith wrote:
> Can't locate Parrot/OpLib/core.pm in @INC (@INC contains:
This file is supposted to be autogenerated from core.ops, however the
dependincies in the makefile are wrong.
For now, you can do a "make Parrot/OpLib/core.pm", and it should do the
Right Thin
On Tue, 16 Oct 2001, Dan Sugalski wrote:
> We sort of do, if you treat string contents as a buffer of bytes rather
> than characters. Raw memory sort of lives a step below the interpreter at
> the moment, though I can see uses for that not being the case.
Hm. So you're thinking that allocating n
On Tue, 2001-10-16 at 13:04, Alex Gough wrote:
> On Tue, 16 Oct 2001, Alex Gough wrote:
> > On 16 Oct 2001, Brian Wheeler wrote:
> >
> > > I'm getting some weird results when using substr. Here's my test
> > > program:
> >
> > It's probably something wrong with the constant table or the assembl
On Tue, 16 Oct 2001, Alex Gough wrote:
> On 16 Oct 2001, Brian Wheeler wrote:
>
> > I'm getting some weird results when using substr. Here's my test
> > program:
>
> It's probably something wrong with the constant table or the assembly
> phase, if the script is changed so that S1 is set to "-",
On 16 Oct 2001, Brian Wheeler wrote:
> I'm getting some weird results when using substr. Here's my test
> program:
It's probably something wrong with the constant table or the assembly
phase, if the script is changed so that S1 is set to "-", say, it does
more what I expect.
set S0
At 11:13 AM 10/14/2001 -0400, James Mastros wrote:
>Hey all.
> I've noticed that we don't seem to have the concept of RAM.
We sort of do, if you treat string contents as a buffer of bytes rather
than characters. Raw memory sort of lives a step below the interpreter at
the moment, though I can
At 05:21 PM 10/16/2001 +0100, Tom Hughes wrote:
>In message <[EMAIL PROTECTED]>
> Simon Cozens <[EMAIL PROTECTED]> wrote:
>
> > Commit it for now, but I'd really, really love it if we could
> > automate this sort of thing.
>
>I have a plan to semi-automate it which I nearly implemted the
I just subscribed and downloaded the latest parrot build. I did do a Google
search on
the following messages before posting here, but no results, and I did read
the README :)
Following the install instructions
make test_prog finishes fine.
Then
[root@cpan parrot]# make test
...
In message <[EMAIL PROTECTED]>
Simon Cozens <[EMAIL PROTECTED]> wrote:
> Commit it for now, but I'd really, really love it if we could
> automate this sort of thing.
I have a plan to semi-automate it which I nearly implemted the other
day but didn't get around to. Basically the idea is
I'm getting some weird results when using substr. Here's my test
program:
set S0,"Hello world"
print "Arg to Reverse: "
print S0
print "\n"
set S1,""
set S2,""
length I0,S0
dec I0
$loop: substr S2,S0,I0,1
Commit it for now, but I'd really, really love it if we could
automate this sort of thing.
--
..you could spend *all day* customizing the title bar. Believe me. I
speak from experience."
(By Matt Welsh)
Heheh, I should read all of my mail before I send new ones. I'll commit
it shortly.
Brian
On Tue, 2001-10-16 at 10:36, Dan Sugalski wrote:
> At 10:04 AM 10/16/2001 -0500, Brian Wheeler wrote:
> >Thoughts?
>
> Go for it. This sort of thing's just fine. I know I made a "NO NEW OPCODES
> WITHOUT
Here's a patch which adds the 'missing' opcodes from the earlier email.
It also adds the 3 arg variant of concat.
Dan/Simon/Anyone, if it seems ok, I'll commit it, but since it adds 52
op variants, I wasn't sure if it would be ok.
Brian
Ops follow
+AUTO_OP add(i, i, ic) {
+AUTO_OP add(n,
At 10:04 AM 10/16/2001 -0500, Brian Wheeler wrote:
>Thoughts?
Go for it. This sort of thing's just fine. I know I made a "NO NEW OPCODES
WITHOUT CLEARANCE" statement a while ago, but that really needs clarification.
The thing that's a no-no is new 'high-level' opcodes--basically anything
that'
After writing a couple of library functions, I realized that we have to
do alot of data shuffling to do common tasks. Reserving a register to
hold 0 or 1 and/or filling up registers with constants just takes up cpu
time and could better be handled if the opcodes took constants directly
as well a
On Mon, 2001-10-15 at 21:12, Dan Sugalski wrote:
> On 15 Oct 2001, Brian Wheeler wrote:
>
> > With the addition of clone, I started writing some generic routines
> > which might be useful (index,lc,uc,reverse,abs,tr,etc)...and I came
> > across some weirdness:
> >
> > doing:
> >save S0
> >
Jason--
> sys/time.h must be included for struct timeval for gettimeofday().
I just added this line to the preamble of core.ops.
> The current
> CVS source is not compiling for me without this. Perl's Config module contains
> a symbol for gettimeofday's availability, so it would be possible to
On Mon, 15 Oct 2001 22:12:58 -0400 (EDT), Dan Sugalski wrote:
>>doing:
>> save S0
>> restore S1
>>
>>(since there's no set S1,S0)
>>
>>binds the registers together, so a change to one is a change to
>>both...which doesn't happen on int registers.
>Right. Save on a string register pushes the
41 matches
Mail list logo
