**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@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On Thu, Oct 9, 2008 at 12:54 AM, nicolas cellier <ncellier@ifrance.com> wrote:

http://bugs.squeak.org/view.php?id=6994 deprecate the Service related #startsWith: because single implementor completely redundant with beginsWith:

#startsWith: is not there anymore but it's still used (e.g. by the mercury field in OB) -- Damien Pollet type less, do more [ | ] http://people.untyped.org/damien.pollet

In my image the problems are that OB-Enhancements and OB-Standard each have a sender of #startsWith: OBEnrichedMercuryQuery class>>selectClassForQuery: OBCodeBrowser class>>defaultBrowserClass Replacing them with #beginsWith: fixes the problem. On Mon, Oct 13, 2008 at 4:44 PM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:

Sequenceable>>startsWith: start | comp | self size < start size ifTrue: [^ false]. comp := true. (self first: start size) with: start do: [:ea :ea2 | ea = ea2 ifFalse: [comp := false]]. ^ comp

And indeed with should use deprecate:...

Stef

On Oct 13, 2008, at 4:34 PM, Damien Pollet wrote:

...

On Thu, Oct 9, 2008 at 12:54 AM, nicolas cellier <ncellier@ifrance.com> wrote:

...

http://bugs.squeak.org/view.php?id=6994 deprecate the Service related #startsWith: because single implementor completely redundant with beginsWith:

#startsWith: is not there anymore but it's still used (e.g. by the mercury field in OB)

-- Damien Pollet type less, do more [ | ] http://people.untyped.org/damien.pollet

_______________________________________________ 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

-- Damien Pollet type less, do more [ | ] http://people.untyped.org/damien.pollet

Thanks, Nicolas! These fixes will appear in update 10095. It would be helpful if you could also create a report in the issue tracker and post the comments there too. This helps us to better track changes. I created the following issue for this mail: http://code.google.com/p/pharo/issues/detail?id=245 Cheers, Adrian On Oct 9, 2008, at 00:54 , 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@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

On Oct 13, 2008, at 22:40 , Stéphane Ducasse wrote:

But I harvested them in 10093 :)

Oh, really? When I merged the slice from the inbox I think I saw changes and I don't think your log mentions these fixes. Adrian

Stef

On Oct 13, 2008, at 10:19 PM, Adrian Lienhard wrote:

...

Thanks, Nicolas!

These fixes will appear in update 10095.

It would be helpful if you could also create a report in the issue tracker and post the comments there too. This helps us to better track changes. I created the following issue for this mail: http://code.google.com/p/pharo/issues/detail?id=245

Cheers, Adrian

On Oct 9, 2008, at 00:54 , 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@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

_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Stéphane Ducasse <stephane.ducasse@...> writes:

Tx. We should use more often deprecate: because this is a cheap way to support evolution without impacting directly clients. Stef

On Oct 13, 2008, at 9:01 PM, Damien Pollet wrote:

...

In my image the problems are that OB-Enhancements and OB-Standard each have a sender of #startsWith: OBEnrichedMercuryQuery class>>selectClassForQuery: OBCodeBrowser class>>defaultBrowserClass

Replacing them with #beginsWith: fixes the problem.

Please note that original Collections-nice.111 has the deprecated method (at least my local copy of it). SequenceableCollection>>startsWith: aCollection self deprecated: 'Use beginsWith:.'. ^self beginsWith: aCollection Since I replaced two sole senders in Base image (Service-Base-nice.38), maybe Stef was overzealous ? PS: Sorry if I repost, but first attempt did not reach the list...

On Oc

Please note that original Collections-nice.111 has the deprecated method (at least my local copy of it).

SequenceableCollection>>startsWith: aCollection

self deprecated: 'Use beginsWith:.'. ^self beginsWith: aCollection

Since I replaced two sole senders in Base image (Service-Base-nice. 38), maybe Stef was overzealous ?

No I did not changed anything. So this is strange because we should not loose code like that.