[Pharo-project] Closure compiler
On 4/8/09, Marcus Denker <denker@acm.org> wrote:
On 07.04.2009, at 16:31, Alexandr Batalshikov wrote:
Hi,
I see in pharo0.1-10268dev09.04.1.image:
Compiler class>>closureParserClass ^self error: 'not installed'.
That is right? How to install "closureParserClass"?
these are leftovers from the NewCompiler ClosureCompiler. We will clean this up soon.
What is the story with the NewCompiler? Does Pharo include it? Or are you moving to Eliot's modification of the old compiler with closure support? Gulik. -- http://gulik.pbwiki.com/
On 07.04.2009, at 21:50, Michael van der Gulik wrote:
On 4/8/09, Marcus Denker <denker@acm.org> wrote:
On 07.04.2009, at 16:31, Alexandr Batalshikov wrote:
Hi,
I see in pharo0.1-10268dev09.04.1.image:
Compiler class>>closureParserClass ^self error: 'not installed'.
That is right? How to install "closureParserClass"?
these are leftovers from the NewCompiler ClosureCompiler. We will clean this up soon.
What is the story with the NewCompiler? Does Pharo include it? Or are you moving to Eliot's modification of the old compiler with closure support?
We integrated Eliot's work into Pharo for now. The actual closure realization is different but similar, so there is no reason to not to use his. The old "ClosureCompiler" then will be renamed "NewCompiler" and fixed to generate code for Eliot's closures. The NewCompiler is a bit more modular/easier to understand and butcher, we used in for many experiments already and we will definitly port it. Marcus -- Marcus Denker -- denker@acm.org http://www.marcusdenker.de
On 4/9/09, Marcus Denker <denker@acm.org> wrote:
On 07.04.2009, at 21:50, Michael van der Gulik wrote:
On 4/8/09, Marcus Denker <denker@acm.org> wrote:
On 07.04.2009, at 16:31, Alexandr Batalshikov wrote:
Hi,
I see in pharo0.1-10268dev09.04.1.image:
Compiler class>>closureParserClass ^self error: 'not installed'.
That is right? How to install "closureParserClass"?
these are leftovers from the NewCompiler ClosureCompiler. We will clean this up soon.
What is the story with the NewCompiler? Does Pharo include it? Or are you moving to Eliot's modification of the old compiler with closure support?
We integrated Eliot's work into Pharo for now. The actual closure realization is different but similar, so there is no reason to not to use his. The old "ClosureCompiler" then will be renamed "NewCompiler" and fixed to generate code for Eliot's closures.
The NewCompiler is a bit more modular/easier to understand and butcher, we used in for many experiments already and we will definitly port it.
I wonder if it is worth splitting the NewCompiler (which really does need a better name) into two packages: the front end and the back end. Later this year (or next) I plan to do some hacking on the NewCompiler for integration into SecureSqueak. I'll post something on the NewCompiler mailing list when I get that far; I'll try to contribute stuff back to the NewCompiler project so hopefully they'll percolate through to Pharo, Squeak and other users. Gulik. -- http://gulik.pbwiki.com/
On 08.04.2009, at 17:40, Michael van der Gulik wrote:
I wonder if it is worth splitting the NewCompiler (which really does need a better name) into two packages: the front end and the back end.
New name: yes. So for packages, we have NewParser, AST and the rest. Where would we split? IRBuilder would be the best candidate, I'd say.
Later this year (or next) I plan to do some hacking on the NewCompiler for integration into SecureSqueak. I'll post something on the NewCompiler mailing list when I get that far; I'll try to contribute stuff back to the NewCompiler project so hopefully they'll percolate through to Pharo, Squeak and other users.
Cool! Marcus -- Marcus Denker -- denker@acm.org http://www.marcusdenker.de
(note: cross-posted on newcompiler and pharo-project lists) On 4/9/09, Marcus Denker <denker@acm.org> wrote:
On 08.04.2009, at 17:40, Michael van der Gulik wrote:
I wonder if it is worth splitting the NewCompiler (which really does need a better name) into two packages: the front end and the back end.
New name: yes.
Umm... HannanCompiler (after the original author?).
So for packages, we have NewParser, AST and the rest. Where would we split? IRBuilder would be the best candidate, I'd say.
I'll get back to you on this when I have stuff to contribute. I'd like to separate out the IR classes so I can use them without needing to install the rest of the compiler. Gulik. -- http://gulik.pbwiki.com/
On 09.04.2009, at 02:49, Stéphane Ducasse wrote:
New name: yes.
Umm... HannanCompiler (after the original author?).
umm..... rot and dead compiler that marcus resurrected
I wonder what happened to Anthony... he back than did work one year full time on the compiler, and than even got a Phd position at a good university in the US. lots of cool ideas... e.g. in the direction of a secure Squeak: http://home.cc.gatech.edu/tony/uploads/32/SharedSmalltalk.htm But he seems to have disappeared... Marcus -- Marcus Denker -- denker@acm.org http://www.marcusdenker.de
On 4/9/09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
New name: yes.
Umm... HannanCompiler (after the original author?).
umm..... rot and dead compiler that marcus resurrected
That was an inappropriate and unhelpful comment. Gulik. -- http://gulik.pbwiki.com/
Oh yes it was appropriate I did not like simply the fact that you could imply that new compiler would be better named HannanCompiler since people could understand that we were stealing things by not mentioning his name. Now of course you may not have implied that and you may not like my comment. Still my point is clear. I'm greatfull to marcus to have resurrected and maintained this software over the years. Stef On Apr 11, 2009, at 2:32 AM, Michael van der Gulik wrote:
On 4/9/09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
New name: yes.
Umm... HannanCompiler (after the original author?).
umm..... rot and dead compiler that marcus resurrected
That was an inappropriate and unhelpful comment.
Gulik.
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On 4/12/09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
Oh yes it was appropriate I did not like simply the fact that you could imply that new compiler would be better named HannanCompiler since people could understand that we were stealing things by not mentioning his name. Now of course you may not have implied that and you may not like my comment. Still my point is clear. I'm greatfull to marcus to have resurrected and maintained this software over the years.
I certainly didn't mean any disrespect or lack of admiration for what Marcus has done. It was just a name to avoid ambiguity. Do you have another suggestion for a name? Gulik. -- http://gulik.pbwiki.com/
OK context taken into account. Suggestion Pluggable Compiler Layered Compiler OpCode: the Smalltalk Open Compiler :) Stef On Apr 12, 2009, at 12:30 AM, Michael van der Gulik wrote:
On 4/12/09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
Oh yes it was appropriate I did not like simply the fact that you could imply that new compiler would be better named HannanCompiler since people could understand that we were stealing things by not mentioning his name. Now of course you may not have implied that and you may not like my comment. Still my point is clear. I'm greatfull to marcus to have resurrected and maintained this software over the years.
I certainly didn't mean any disrespect or lack of admiration for what Marcus has done. It was just a name to avoid ambiguity.
Do you have another suggestion for a name?
Gulik.
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
What about a simple (Squeak Smalltalk) compiler version 2 or compiler 2009? Some time in the future even this "new compiler" will probably an "old compiler" if Squeak survives long enough. Andreas Am 12.04.2009 um 10:50 schrieb Stéphane Ducasse:
OK context taken into account.
Suggestion
Pluggable Compiler Layered Compiler
OpCode: the Smalltalk Open Compiler
:)
Stef
On Apr 12, 2009, at 12:30 AM, Michael van der Gulik wrote:
On 4/12/09, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
Oh yes it was appropriate I did not like simply the fact that you could imply that new compiler would be better named HannanCompiler since people could understand that we were stealing things by not mentioning his name. Now of course you may not have implied that and you may not like my comment. Still my point is clear. I'm greatfull to marcus to have resurrected and maintained this software over the years.
I certainly didn't mean any disrespect or lack of admiration for what Marcus has done. It was just a name to avoid ambiguity.
Do you have another suggestion for a name?
Gulik.
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
participants (4)
-
Andreas Wacknitz -
Marcus Denker -
Michael van der Gulik -
Stéphane Ducasse