On 12/21/12, Stéphane Rollandin <[hidden email]> wrote:
.... > Now the issue of where to find code is a real one; in my 10+ years of > experience with Squeak it has always been. I see this as a reflect of > the nature of Squeak as an extraordinary playground for prototyping > stuff. Code is wild in Squeak: both inside the image, and outside. ... > My own contribution for Squeak simply have > entries in SqueakMap and are available from my site either as sar files > or ready-to-go images: > > http://www.zogotounga.net/comp/eindex.html Stéphane Thank you for "re-announcing" your code repository. Maybe you should do that semi-annually... BTW what about LispKit and Squeak 4.4? http://www.zogotounga.net/comp/squeak/lispkit.htm Does it load and work fine? --Hannes |
> BTW what about LispKit and Squeak 4.4?
> http://www.zogotounga.net/comp/squeak/lispkit.htm > Does it load and work fine? It didn't actually; I just uploaded a new version that seems ok. Prolog does not load in a 4.4 image though; I'll have a look at that ASAP. Stef |
Thank you Stéphane for updating the SqueakMap entry for the Lisp-Kit.
My internet connection has gone slow. Take your time and I'll do a load-test for both. --Hannes On 12/21/12, Stéphane Rollandin <[hidden email]> wrote: >> BTW what about LispKit and Squeak 4.4? >> http://www.zogotounga.net/comp/squeak/lispkit.htm >> Does it load and work fine? > > It didn't actually; I just uploaded a new version that seems ok. > > Prolog does not load in a 4.4 image though; I'll have a look at that ASAP. > > Stef > > > |
Prolog is tricky. Did something change recently regarding #compilerClass
and its usage ? Stef |
> Prolog is tricky. Did something change recently regarding #compilerClass
> and its usage ? Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser subclass, so if something changed in Parser that may be the reason it does not load. Stef |
First thing, you should replace (30 asCharacter) with DoItCharacter (a
class variable) in #xLetter I can't remember if there is other change with potential impact... Nicolas 2012/12/21 Stéphane Rollandin <[hidden email]>: >> Prolog is tricky. Did something change recently regarding #compilerClass >> and its usage ? > > > Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser > subclass, so if something changed in Parser that may be the reason it does > not load. > > Stef > > |
Then, (type _ typeTable at: c ifAbsent: [#xLetter]) should be replaced
with (type _ self typeTableAt: c), otherwise it won't handle DoItCharacter Nicolas 2012/12/21 Nicolas Cellier <[hidden email]>: > First thing, you should replace (30 asCharacter) with DoItCharacter (a > class variable) in #xLetter > I can't remember if there is other change with potential impact... > > Nicolas > > 2012/12/21 Stéphane Rollandin <[hidden email]>: >>> Prolog is tricky. Did something change recently regarding #compilerClass >>> and its usage ? >> >> >> Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser >> subclass, so if something changed in Parser that may be the reason it does >> not load. >> >> Stef >> >> |
it loads with attached definition of xLetter
2012/12/21 Nicolas Cellier <[hidden email]>: > Then, (type _ typeTable at: c ifAbsent: [#xLetter]) should be replaced > with (type _ self typeTableAt: c), otherwise it won't handle > DoItCharacter > > Nicolas > > 2012/12/21 Nicolas Cellier <[hidden email]>: >> First thing, you should replace (30 asCharacter) with DoItCharacter (a >> class variable) in #xLetter >> I can't remember if there is other change with potential impact... >> >> Nicolas >> >> 2012/12/21 Stéphane Rollandin <[hidden email]>: >>>> Prolog is tricky. Did something change recently regarding #compilerClass >>>> and its usage ? >>> >>> >>> Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser >>> subclass, so if something changed in Parser that may be the reason it does >>> not load. >>> >>> Stef >>> >>> LParser-xLetter.st (1K) Download Attachment |
In reply to this post by Nicolas Cellier
On 21 December 2012 12:39, Nicolas Cellier
<[hidden email]> wrote: > Then, (type _ typeTable at: c ifAbsent: [#xLetter]) should be replaced > with (type _ self typeTableAt: c), otherwise it won't handle > DoItCharacter Do you mean rather: (type := self typeTableAt: c) ? :) frank > Nicolas > > 2012/12/21 Nicolas Cellier <[hidden email]>: >> First thing, you should replace (30 asCharacter) with DoItCharacter (a >> class variable) in #xLetter >> I can't remember if there is other change with potential impact... >> >> Nicolas >> >> 2012/12/21 Stéphane Rollandin <[hidden email]>: >>>> Prolog is tricky. Did something change recently regarding #compilerClass >>>> and its usage ? >>> >>> >>> Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser >>> subclass, so if something changed in Parser that may be the reason it does >>> not load. >>> >>> Stef >>> >>> > |
I let Stephane use _ if he wants too, with proper Preferences, that should work.
The rationale for DoItCharacter change can be found here http://forum.world.st/The-Trunk-Compiler-nice-223-mcz-td4373145.html Or simply open Monticello browser, select Compiler and browse history, Stephane, as you can see, an SCM has advantages too. Nicolas 2012/12/21 Frank Shearar <[hidden email]>: > On 21 December 2012 12:39, Nicolas Cellier > <[hidden email]> wrote: >> Then, (type _ typeTable at: c ifAbsent: [#xLetter]) should be replaced >> with (type _ self typeTableAt: c), otherwise it won't handle >> DoItCharacter > > Do you mean rather: (type := self typeTableAt: c) ? > > :) > > frank > >> Nicolas >> >> 2012/12/21 Nicolas Cellier <[hidden email]>: >>> First thing, you should replace (30 asCharacter) with DoItCharacter (a >>> class variable) in #xLetter >>> I can't remember if there is other change with potential impact... >>> >>> Nicolas >>> >>> 2012/12/21 Stéphane Rollandin <[hidden email]>: >>>>> Prolog is tricky. Did something change recently regarding #compilerClass >>>>> and its usage ? >>>> >>>> >>>> Or did Parser changed between 4.3 and 4.4 ? Prolog has its own Parser >>>> subclass, so if something changed in Parser that may be the reason it does >>>> not load. >>>> >>>> Stef >>>> >>>> >> > |
In reply to this post by Nicolas Cellier
Ok, fixed
I uploaded on SqueakMap and on my site a new version of Prolog, which loads and works on 4.4 Stef |
In reply to this post by Nicolas Cellier
> I let Stephane use _ if he wants too, with proper Preferences, that should work.
All my := are _ , over thousands of lines of code... Still no tool for performing the replacement in a 100 % guaranteed safe way ? Stef |
On 21 December 2012 14:29, Stéphane Rollandin <[hidden email]> wrote:
>> I let Stephane use _ if he wants too, with proper Preferences, that should >> work. > > > All my := are _ , over thousands of lines of code... Still no tool for > performing the replacement in a 100 % guaranteed safe way ? > It exists in Pharo for years , see FixUnderscores package. I think you can use it in squeak without much changes. > Stef -- Best regards, Igor Stasenko. |
>> All my := are _ , over thousands of lines of code... Still no tool for
>> performing the replacement in a 100 % guaranteed safe way ? >> > > It exists in Pharo for years , see FixUnderscores package. > > I think you can use it in squeak without much changes. Found it on SqueakMap, thanks. Stef |
In reply to this post by Stéphane Rollandin
On Fri, Dec 21, 2012 at 3:42 AM, Stéphane Rollandin <[hidden email]> wrote: Prolog is tricky. Did something change recently regarding #compilerClass and its usage ? Yes. It got replaced with newCompiler, and parserClass by newParser. These are marked improvements because they allow answering an initialized object, which gives much more scope for difference.
best, Eliot
|
2012/12/21 Eliot Miranda <[hidden email]>:
> > > > On Fri, Dec 21, 2012 at 3:42 AM, Stéphane Rollandin <[hidden email]> > wrote: >> >> Prolog is tricky. Did something change recently regarding #compilerClass >> and its usage ? > > > Yes. It got replaced with newCompiler, and parserClass by newParser. These > are marked improvements because they allow answering an initialized object, > which gives much more scope for difference. > But for a subclass that just refine parserClass, this change should be transparent. Nicolas >> >> >> Stef >> > > > > -- > best, > Eliot > > > |
Free forum by Nabble | Edit this page |