- Pharo-dev - lists.pharo.org

[Pharo-project] Some little syntax fix.

by Mathieu Suen

Hi, I have some little syntax fix. This will let the newcompiler recompile the image. Mth

Oct. 9, 2008

Re: [Pharo-project] Monticello: an idealist question

by Bill Schwab

Thanks for the reply. Please do not feel attacked in any way by this, but you have beautifully illustrated my point: I am not asking to put method in multiple *packages*, I am asking to free categories to do what they *should* do, which is tag methods in ways I find useful. PackageInfo (not to denigrate it), has completely hijacked the concept - actually, the Squeak community failed to add state to allow it to operate independently of categories. The uses of multiple categories are probably obvious, but they can be as simple as marking something as, for example, being a helper for the FFT, and specifically applying to the frequency domain. Dolphin visually distinguishes public and private methods, so I tend to be generous with the private category as I find it helps me when I look at a class with many methods. It is difficult to miss something you have never had, but trust me when I say that you are missing out. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029 >>> tapplek(a)gmail.com 10/08/08 10:33 PM >>> On Wed, Oct 08, 2008 at 07:52:29PM -0400, Bill Schwab wrote: > Colin, > > Among many useful things in Dolphin is an ability to assign multiple > categories to any given method; Squeak and Pharo allow only one > category, and it promptly gets chewed up by PackageInfo. Multiple > categories would be a nice addition to Squeak and Pharo. > > How hard would it be to change Monticello to take its input not from > PackageInfo and its (pardon me) abuse of category names, but from > reifications of the categories? The new objects would form the basis > for packaging system, and free method categories to do what they do > best: aid in browsing and understanding code. You could change PackageInfo to use the new scheme easily, I'm sure, or you could use Monticello2, which adds another abstraction layer atop PackageInfo (slices). And if Craig ever releases Spoon, we'll have another contender for version management (Naiad) PackageInfo is not that complicated. I was able to rewrite its image scanning code to be about 8x faster in about 40 lines of code, spread across 8 methods Making the browser nicely cope with whatever scheme you invent will be the real challenge. Can you give me an example of why you want one method in two packages? Monticello can handle that by using overrides, but it is a different version in each package. -- Matthew Fulmer -- http://mtfulmer.wordpress.com/ _______________________________________________ Pharo-project mailing list Pharo-project(a)lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 9, 2008

Re: [Pharo-project] Monticello: an idealist question

by Damien Pollet

On Thu, Oct 9, 2008 at 4:33 AM, Matthew Fulmer <tapplek(a)gmail.com> wrote: > PackageInfo is not that complicated. Some might even say it's rather simplistic >:] It would make sense to have real objects reify the code organization. > Can you give me an example of why you want one method in two > packages? Monticello can handle that by using overrides, but it > is a different version in each package. I guess he meant method categories (protocols) ? -- Damien Pollet type less, do more [ | ] http://people.untyped.org/damien.pollet

Oct. 9, 2008

[Pharo-project] Pharo one click on Win32 - VM fix available + new Freetype plugin

by Torsten Bergmann

The Pharo one click VM for Win32 still has the full screen problem, which is already fixed for Squeak by Andreas with a newer VM. See http://bugs.squeak.org/view.php?id=6818 for details and the new sources. There is also a new build of the external Freetype Plugin. See http://lists.squeakfoundation.org/pipermail/squeak-dev/2008-October/131889.… Thx Torsten -- Psssst! Schon vom neuen GMX MultiMessenger gehÃ¶rt? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

Oct. 9, 2008

Re: [Pharo-project] SLICE-Mantis-Collection-Fix-mining-nice-2

by StÃ©phane Ducasse

**Thanks** I will harvest that stef On Oct 9, 2008, at 12:54 AM, nicolas cellier wrote: > I have uploaded a SLICE in PharoInbox gathering some more or less > important collection fixes from mantis. > > > GATHER SEVERAL COLLECTION RELATED MANTIS TESTS AND FIXES > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7172 > > CharacterSetComplement has spoiled previous byteArrayMap optimization. > Using #byteArrayMap is an optimization per se, possibly tailored for > tight loops. > Taking too much time to compute this map is spoiling efficiency. > It is necessary to cache this map for restoring efficiency to the > level preceding introduction of CharacterSetComplement. > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6998 > > ((String new: 1) at: 1 put: $a) = $a -> true. > ((WideString new: 1) at: 1 put: $a) = $a -> false. > > WideString at:put: should return the put object like any other > collection > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=5331 > http://bugs.squeak.org/view.php?id=6366 > > Lots of tests and few patches for WideString not behaving correctly. > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 > > Patches for Interval indexOf: and includes: > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6994 > > Authorize this: > (#(1 2 3) as: ByteArray) as: Interval > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7121 (ar) > > Patches for OrderedCollection add:after: add:before: > Try this one to be convinced for the necessity: > | oc | > oc := #(1 2 3 4) asOrderedCollection. > (oc first:3) do: [:e | oc remove: e]. > oc add: 0 beforeIndex: -2. > ^oc > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6977 > incorrect use of super > (super basicNew in #new: dating from some 3.9 changes) > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6778 > SkipList copy is shallow > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6994 > deprecate the Service related #startsWith: because single > implementor completely redundant with beginsWith: > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 9, 2008

Re: [Pharo-project] SLICE-Mantis-Collection-Fix-mining-nice-2

by StÃ©phane Ducasse

**Thanks** I will harvest that stef On Oct 9, 2008, at 12:54 AM, nicolas cellier wrote: > I have uploaded a SLICE in PharoInbox gathering some more or less > important collection fixes from mantis. > > > GATHER SEVERAL COLLECTION RELATED MANTIS TESTS AND FIXES > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7172 > > CharacterSetComplement has spoiled previous byteArrayMap optimization. > Using #byteArrayMap is an optimization per se, possibly tailored for > tight loops. > Taking too much time to compute this map is spoiling efficiency. > It is necessary to cache this map for restoring efficiency to the > level preceding introduction of CharacterSetComplement. > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6998 > > ((String new: 1) at: 1 put: $a) = $a -> true. > ((WideString new: 1) at: 1 put: $a) = $a -> false. > > WideString at:put: should return the put object like any other > collection > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=5331 > http://bugs.squeak.org/view.php?id=6366 > > Lots of tests and few patches for WideString not behaving correctly. > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7180 +6455 +1603 +1602 > > Patches for Interval indexOf: and includes: > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6994 > > Authorize this: > (#(1 2 3) as: ByteArray) as: Interval > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=7121 (ar) > > Patches for OrderedCollection add:after: add:before: > Try this one to be convinced for the necessity: > | oc | > oc := #(1 2 3 4) asOrderedCollection. > (oc first:3) do: [:e | oc remove: e]. > oc add: 0 beforeIndex: -2. > ^oc > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6977 > incorrect use of super > (super basicNew in #new: dating from some 3.9 changes) > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6778 > SkipList copy is shallow > > ---------------------------------------- > > http://bugs.squeak.org/view.php?id=6994 > deprecate the Service related #startsWith: because single > implementor completely redundant with beginsWith: > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 9, 2008

Re: [Pharo-project] Polymorph

by Damien Cassou

On Wed, Oct 8, 2008 at 9:54 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > why don't you provide a PolymorphAll package which lists all the other as > required. This is a good solution for developers but not for the universe. In the Universe, you have to declare all subpackages. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu)

Oct. 9, 2008

Re: [Pharo-project] how to register a tool in the open menu?

by Damien Cassou

On Wed, Oct 8, 2008 at 8:07 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > Damien could you add the MethodFinder to the tools you add to the oneClick > image. As soon as you tell me it's ready and its repository. -- Damien Cassou Peter von der AhÃ©: Â«I'm beginning to see why Gilad wished us good luckÂ». (http://blogs.sun.com/ahe/entry/override_snafu)

Oct. 9, 2008

Re: [Pharo-project] Monticello: an idealist question

by Matthew Fulmer

On Wed, Oct 08, 2008 at 07:52:29PM -0400, Bill Schwab wrote: > Colin, > > Among many useful things in Dolphin is an ability to assign multiple > categories to any given method; Squeak and Pharo allow only one > category, and it promptly gets chewed up by PackageInfo. Multiple > categories would be a nice addition to Squeak and Pharo. > > How hard would it be to change Monticello to take its input not from > PackageInfo and its (pardon me) abuse of category names, but from > reifications of the categories? The new objects would form the basis > for packaging system, and free method categories to do what they do > best: aid in browsing and understanding code. You could change PackageInfo to use the new scheme easily, I'm sure, or you could use Monticello2, which adds another abstraction layer atop PackageInfo (slices). And if Craig ever releases Spoon, we'll have another contender for version management (Naiad) PackageInfo is not that complicated. I was able to rewrite its image scanning code to be about 8x faster in about 40 lines of code, spread across 8 methods Making the browser nicely cope with whatever scheme you invent will be the real challenge. Can you give me an example of why you want one method in two packages? Monticello can handle that by using overrides, but it is a different version in each package. -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

Oct. 9, 2008

[Pharo-project] Monticello: an idealist question

by Bill Schwab

Colin, Among many useful things in Dolphin is an ability to assign multiple categories to any given method; Squeak and Pharo allow only one category, and it promptly gets chewed up by PackageInfo. Multiple categories would be a nice addition to Squeak and Pharo. How hard would it be to change Monticello to take its input not from PackageInfo and its (pardon me) abuse of category names, but from reifications of the categories? The new objects would form the basis for packaging system, and free method categories to do what they do best: aid in browsing and understanding code. In transition, a future version of Monticello could create the needed objects using the same rules as PackageInfo (or any time it is instructed to do so). Comments? Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 273-6785 FAX: (352) 392-7029

Oct. 8, 2008