Squeak-dev March 2015

squeak-dev@lists.squeakfoundation.org

40 participants
315 discussions

The Trunk: Morphic-mt.767.mcz
by commits＠source.squeak.org 01 Mar '15

01 Mar '15

Marcel Taeumel uploaded a new version of Morphic to project The Trunk: http://source.squeak.org/trunk/Morphic-mt.767.mcz ==================== Summary ==================== Name: Morphic-mt.767 Author: mt Time: 1 March 2015, 5:41:07.646 pm UUID: aeb8e33f-7b90-7545-b3fa-b83eb88cad79 Ancestors: Morphic-mt.766 Make TransformMorph use #clipSubmorphs: as specified in Morph. Having this, TransforMorphs can have a #layoutPolicy: again. #layoutBounds overwritten to respect local coordinate system. =============== Diff against Morphic-mt.766 =============== Item was removed: - ----- Method: TransformMorph>>fullBounds (in category 'layout') ----- - fullBounds - "Overridden to clip submorph hit detection to my bounds." - "It might be better to override doLayoutIn:, and remove this method" - - fullBounds ifNotNil:[^ fullBounds]. - fullBounds := bounds. - submorphs do: [:m| m ownerChanged]. - ^ fullBounds! Item was changed: ----- Method: TransformMorph>>initialize (in category 'initialization') ----- initialize "initialize the state of the receiver" super initialize. "" smoothing := 1. + transform := MorphicTransform identity. + self clipSubmorphs: true.! - transform := MorphicTransform identity! Item was added: + ----- Method: TransformMorph>>layoutBounds (in category 'layout') ----- + layoutBounds + + ^ (transform globalBoundsToLocal: super layoutBounds) truncated + !

1 0

last update 299 might be broken
by Nicolas Cellier 01 Mar '15

01 Mar '15

Hi, it seems that last update is loading Graphics first and thus removing CornerRounder before loading Morphic. Alas, the old Morphic might still be using CornerRounder. In some old images still using some rounded things it happened that the update process failed with unresponsive image (recursive error or something...). I disabled all the rounded* Preferences, but it does not seem to be enough (I think the progress dialog itself must be using rounded corners or some thing like that...). Can we fix that? (for example just delay inclusion of latest Graphics - that is exclude it from the latest update)

3 5

The Trunk: Tools-mt.535.mcz
by commits＠source.squeak.org 01 Mar '15

01 Mar '15

Marcel Taeumel uploaded a new version of Tools to project The Trunk: http://source.squeak.org/trunk/Tools-mt.535.mcz ==================== Summary ==================== Name: Tools-mt.535 Author: mt Time: 1 March 2015, 10:46:38.721 am UUID: 5406c378-35cd-1d43-aa16-abb5ec8b11e2 Ancestors: Tools-mt.534 New preference for auto-sorting the message category list. Disabled by default. =============== Diff against Tools-mt.534 =============== Item was changed: CodeHolder subclass: #Browser instanceVariableNames: 'environment systemOrganizer classOrganizer metaClassOrganizer editSelection metaClassIndicated selectedSystemCategory selectedClassName selectedMessageName selectedMessageCategoryName' + classVariableNames: 'ListClassesHierarchically RecentClasses SortMessageCategoriesAlphabetically' - classVariableNames: 'ListClassesHierarchically RecentClasses' poolDictionaries: '' category: 'Tools-Browser'! !Browser commentStamp: 'cwp 12/27/2012 11:09' prior: 0! I represent a query path into the class descriptions, the software of the system.! Item was added: + ----- Method: Browser class>>sortMessageCategoriesAlphabetically (in category 'preferences') ----- + sortMessageCategoriesAlphabetically + + <preference: 'Sort Message Categories Alphabetically' + categoryList: #(browsing Tools) + description: 'When enabled, the message category list in the browser is ordered by alphabet. No need to ''alphabetize'' this list manually via the context menu.' + type: #Boolean> + + ^ SortMessageCategoriesAlphabetically ifNil: [false] + ! Item was added: + ----- Method: Browser class>>sortMessageCategoriesAlphabetically: (in category 'preferences') ----- + sortMessageCategoriesAlphabetically: aBoolean + + SortMessageCategoriesAlphabetically := aBoolean. + ! Item was changed: ----- Method: Browser>>messageCategoryList (in category 'message category list') ----- messageCategoryList "Answer the selected category of messages." + ^ self hasClassSelected + ifFalse: [Array new] + ifTrue: [ {ClassOrganizer allCategory}, + (self class sortMessageCategoriesAlphabetically + ifTrue: [self rawMessageCategoryList sorted] + ifFalse: [self rawMessageCategoryList]) ]! - self hasClassSelected - ifTrue: [^ (Array with: ClassOrganizer allCategory), self classOrMetaClassOrganizer categories sorted] - ifFalse: [^ Array new]!

1 0

AW: Re: [squeak-dev] community decision making (was: The Trunk: Tools-mt.534.mcz)
by Herbert König 01 Mar '15

01 Mar '15

Hi, I voted for sort preferece I have to enable because i have a visual memory and resorting forces me to search. For new code it doesn't matter but once I used a (any) order I must not change it. One vote explained :-) Cheers Herbert (phone) <div>-------- Ursprüngliche Nachricht --------</div><div>Von: "David T. Lewis" <lewis(a)mail.msen.com> </div><div>Datum:28.02.2015 20:26 (GMT+01:00) </div><div>An: ma.chris.m(a)gmail.com,The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org> </div><div>Betreff: Re: [squeak-dev] community decision making (was: The Trunk: Tools-mt.534.mcz) </div><div> </div>On Sat, Feb 28, 2015 at 12:48:14PM -0600, Chris Muller wrote: > > Ha! I just voted, and it looks like the number of different opinions > > exactly matches the number of people who have voted so far. > > > > If we could get one more person to vote, then stop the election, we > > would have a decision ;-) > > > > Joking aside, there is something to be said for taking of poll to find > > out what most people want, without regard for who has the loudest opinion. > > It already is without regard to loudness, the loudest voters still > only get one vote. > > > So folks - if you have an opinion on this topic, please follow the link > > above and click on your preference. > > Voting has been a last resort, not a first resort, of decision-making > by this community. When an issue comes up, the people who care enough > about that issue to step up and present the reasons of their position > to their peers, so that a community consensus is arrived at based on > community-input from those with a vested interest. > > By contrast, some external, anonymous poll in which dozens folks who > don't participate much, don't care that much because they started a > Squeak image 3 times in their whole life, -- those folks having as > much pull as everyone else who is more-vested. > > Besides, how can we know such an external poll is even fair? Could > someone vote multiple times from different computers, for example? > Right here, we know any votes are fair, and by the discussions we know > the *reasons* for the outcome.. Sure, no disagreement. Like most attempts at democratic processes, it probably will not work very well. But sometimes we don't actually know how people are using the system, and asking the broader community for input is one way to get information. Think of the poll as a way to ask for input that you can use to make an informed decision. I don't know if it will work, but it seems worth a try, so let's see what happens. Dave

1 0

Re: [squeak-dev] The Trunk: Kernel-ul.906.mcz
by David T. Lewis 01 Mar '15

01 Mar '15

On Thu, Feb 26, 2015 at 09:03:07PM +0000, commits(a)source.squeak.org wrote: > Levente Uzonyi uploaded a new version of Kernel to project The Trunk: > http://source.squeak.org/trunk/Kernel-ul.906.mcz > > ==================== Summary ==================== > > Name: Kernel-ul.906 > Author: ul > Time: 26 February 2015, 9:22:02.722 pm > UUID: 9fd0ffbc-c774-442a-baa5-5b0e13950856 > Ancestors: Kernel-topa.905 > > Added an implementation of a 30-bit Mersenne Twister (p=521) to Random. All existing Random instances are migrated to the new implementation. > I love seeing things like this entering the image. I remember someone long ago (maybe it was Andreas or Dan?) describing Squeak as a "computer science degree in a box". These updates to Random are a very nice example, with clear comments to explain the complicated bits, and a class comment to give an overview with references to the source of the algorithm. Cool. Some updates to RandomTest may be also needed to match the new implementation. Dave

2 1

Jump to page: