Brian Rice | 5 Oct 17:24 2005

Syntax discussions added to the wiki

Hi, all.

I apologize for the recent silence for the last week or two: I've  
been pre-occupied in my personal and social life, and haven't had  
time to slow down and give Slate my full attention. My current  
independent contracting employment will be over at the end of the  
week, and I expect to work on and finish a few projects for Slate and  
maybe ready a new release for consumption.

In the mean-time, I've summarized our recent discussion about  
assignment idioms, along with other issues about Slate syntax/grammar  
that irk me and preserved them here: http://slate.tunes.org/wiki/ 
Roadmap#Syntax

I'm not going to take action on the assignment idiom just yet, but my  
willingness has not changed, and I'll have this available when we  
have a better environment in which to try it out.

I'll follow up with more soon... thanks.

--
-Brian

Brian Templeton | 8 Oct 08:21 2005

Re: A provocative naming idea for "setter" methods

Bill Sun <billksun <at> yahoo.com> writes:

> That sort of sparked another idea here...
>
> What about:
>
> object1 set:
> object1 get
>
> or
>
> object1 slot1 set:
> object1 slot1 get
>

What about using a macro?

m <at> (Syntax UnaryMessage traits) set: v
[
  Syntax KeywordMessage
    sending: (intern: m selector name ; ':')
    to: { m arguments first. v. }
].

addSlot: #foo.
foo `set: 42.

--

-- 
              bpt <bpt <at> tunes.org> http://bpt.tunes.org/               
  (let((!'(let((!'?))(subst !(intern"?")!)))) (subst !(intern"?")!))  
(Continue reading)

bpt | 10 Oct 08:42 2005

darcs patch: Correct names of WriteBufferStream, ReadBufferStream i...

Mon Oct 10 02:36:52 EDT 2005  bpt <at> tunes.org
  * Correct names of WriteBufferStream, ReadBufferStream in reference manual
Attachment: text/x-darcs-patch, 8 KiB
bpt | 12 Oct 01:57 2005

darcs patch: Partially fix #do: for queues

Tue Oct 11 19:56:53 EDT 2005  bpt <at> tunes.org
  * Partially fix #do: for queues
Attachment: text/x-darcs-patch, 7918 bytes
Brian Rice | 18 Oct 16:31 2005

Site online after maintenance, wiki troubles

Just FYI, there was some web server maintenance recently that  
resulted in some excessive unforeseen downtime. A side-effect seems  
to be that the MySQL daemon that is the backend for the wiki did not  
start up properly. It's being looked into.

--
-Brian

Brian Rice | 18 Oct 23:55 2005

Re: Site online after maintenance, wiki troubles

There was a configuration error that has been resolved; the wiki is  
now online again.

On Oct 18, 2005, at 7:31 AM, Brian Rice wrote:

> Just FYI, there was some web server maintenance recently that  
> resulted in some excessive unforeseen downtime. A side-effect seems  
> to be that the MySQL daemon that is the backend for the wiki did  
> not start up properly. It's being looked into.

--
-Brian

Brian Rice | 19 Oct 03:04 2005

Updates

Hi all,

So, after many weeks, my contract job is over and not soon enough.  
I've been neglecting Slate, but it was a busy project and I gained a  
lot of embedded systems experience, so I think it was rewarding.

Some updates since there last was activity:
- The make install target was re-instated. There is still the problem  
that the image needs to start up in the source directory. I have part  
of a general facility working to fix this.
- Nick Forde added many benchmark tests from the Computer Language  
Shootout. "make benchmark" runs them. Verdict: The Slate VM is quite  
fast; the Slate image, incredibly slow (due to huge layers of message- 
passing overhead - we have a lot more library abstraction than other  
systems, supposed to be elided by dynamic inlining later).
- Bug-fixes for queues, ring buffers, and various other bits.

I am going to crank out a new version soon to get all of the  
interesting changes out into the wild so I can focus on bigger  
features that we really need to move forward - more on that later.

--
-Brian

Paul Dufresne | 24 Oct 19:52 2005

About Slate and GUI

Well, I could have put "Just to say I am still alive" as the subject
line, but
I guess "About Slate and GUI" will be more correct for the follows up I
can expect.

I am not too proud of the work on SDL I made for Slate.
I was having a bad time, to understand the DUIM-morphic base Brian did.

I have been trying to read and use DUIM from OpenDylan. Dylan has mainly
two
implementations: Gwydion Dylan d2c (that compile Dylan to C code), and
the
old Harlequin's implementation that have been open-sourced for about 1
year,
and is now maintained by the same guys as Gwydion Dylan, that have
renamed it
OpenDylan.

Before understanding DUIM sources, I had to learn a bit of Dylan. I was
very
impressed and please to see that Dylan had a semantic not so different
than
Slate. I was surprised to learn that Dylan could be seen as a
functionnal language.
I was even more surprised to see that Dylan have multiple dispatch that
I have learn
about for the first time with Slate.
I ask myself what is the difference between a prototype based language
and a dynamic
(one could add new methods while running) like Dylan. Yes Dylan have
(Continue reading)

Lendvai Attila | 25 Oct 10:40 2005
Picon

String refactoring, alpha updated


Hi!

In main: String was cut into String and ASCIIString. It doesn't mean too
many api changes, currently String new will return ASCIIStrings as
before and probably this api will remain so when the i18n package is
fully integrated except it will return a smarter string implementation.
An important change is that Character is now accessible trough its
appropiate String prototypes; e.g. ASCIICharacter became ASCIIString
Character.

Hopefully I didn't break anything, but if something fails drop a mail to
the list.

- 101

Nick Forde | 25 Oct 11:01 2005
Picon
Picon

Re: String refactoring, alpha updated

$ gmake check

Welcome to Slate: Less talk, more rock!
  Version: 0.3.5
   System: Linux [little-endian]
    Build: optimized

Downloading http://slate.tunes.org/repos/alpha/vm.c
Downloading http://slate.tunes.org/repos/alpha/vm.h
Downloading http://slate.tunes.org/repos/alpha/AutoLoad
Compiling src/mobius/vm/platform/ansifile.c  (optimized)
Compiling src/mobius/vm/platform/boot.c  (optimized)
Compiling src/mobius/vm/platform/extprim.c  (optimized)
src/mobius/vm/platform/extprim.c: In function `applyExternalLibraryPrimitive':
src/mobius/vm/platform/extprim.c:301: `ASCIIStringProto' undeclared (first use in this function)
src/mobius/vm/platform/extprim.c:301: (Each undeclared identifier is reported only once
src/mobius/vm/platform/extprim.c:301: for each function it appears in.)
gmake: *** [src/mobius/vm/platform/extprim.o] Error 1

Nick.

Lendvai Attila wrote:
> Hi!
> 
> In main: String was cut into String and ASCIIString. It doesn't mean too
> many api changes, currently String new will return ASCIIStrings as
> before and probably this api will remain so when the i18n package is
> fully integrated except it will return a smarter string implementation.
> An important change is that Character is now accessible trough its
> appropiate String prototypes; e.g. ASCIICharacter became ASCIIString
(Continue reading)


Gmane