Squeak-dev May 2018

squeak-dev@lists.squeakfoundation.org

44 participants
189 discussions

Proof-of-concept: moving Smalltalk code editing support into SmalltalkEditor
by Tim Johnson 30 Aug '18

30 Aug '18

Hi all, I went on a journey this morning which could have been quixotic, but I believe I actually found and slayed a small windmill. I was a bit annoyed by how autoEnclose (and by extension autoIndent) were in effect for even the most basic of TextMorphs, e.g. those dragged out of the Supplies flap. I was also kind of disappointed that turning on/off the auto-enclose preference affected all text morphs, even those not specifically designated for Smalltalk code editing. I was emboldened when I saw the class comment for TextEditor: "[I] have no specific facilities for editing Smalltalk code. Those are found in SmalltalkEditor." I found myself able to address this by making a few changes: * moving autoEnclose and autoIndent functionality out of TextEditor and into SmalltalkEditor * changing the default editor class for TextMorph to TextEditor * specifying that the #editorClass for TextMorphForEditView should be a SmalltalkEditor * changing PreferenceWizard to recognize these preferences have a new home I'm attaching a changeset to this email and not uploading an .mcz to inbox (sorry). This is because I don't know how to represent changes across many projects in Monticello, and I sort of consider this a proof-of-concept rather than something I think warrants immediate inclusion in the image. Please do let me know what you think about this idea. I also fear my implementation of SmalltalkEditor>>#dispatchOnKeyboardEvent: might be lacking. Thanks, Tim ps - I think a next step of moving Smalltalk editing support into SmalltalkEditor might be to move the call to ToolSet>>#codeCompletionAround:textMorph:keyStroke: out of TextMorph>>#keyStroke: and into its subclass, in TextMorphForEditView>>#keyStroke:. I didn't complete that move and did not quite figure out how the subclass version of that method works.

9 19

Survey: what do you do with Squeak, what do you *want* to do?
by Benoit St-Jean 01 Aug '18

01 Aug '18

Q-What do you use Squeak for? Genetic algorithms, scripting (a LOT), experimenting, Euler Project code, a chess engine, random data generator, cleaning up/transforming crappy data from databases and/or other sources, keeping up with new ideas in Smalltalk. Q-If you don't use Squeak, why not? Lack of a decent GUI builder. And unstability of the base code and/or difficulty to find code/frameworks/applications that load in-one-shot without me having to debug in the first place! Q-If you used Squeak in the past and don't now, what pulled you away? Lack of a decent GUI builder. Lack of a decent Windows support. Q-What does Squeak lack that you think might make you use it for 'regular' development? A decent GUI builder. And a central code repository that wouldn't make loading code a crazy adventure every time. If I want Python code, all I have to do is to visit PyPI. If I want Ruby code, all I have to do is to visit RubyGems. If I want R code, all I have to do is to visit CRAN R. If I want VisualWorks code, all I have to do is to visit Cincom Public Repository... Now, if I want Squeak code, I have to visit SqueakSource, SmalltalkHub, GitHub, SqueakSource3, SqueakMap and use a GitHub tool, or file in some code, or use the package manager, or use Monticello, or use the Package Universe browser, or Gofer, or the SqueakMap Catalog Browser, or... Well, you get the picture. Besides, once I've found what I was looking for, often times the code doesn't load because it was ported to Pharo... or it's never been updated to be compatible with the current Squeak version. Another tool Squeak needs is an image packager (à la VW or VAST or Dolphin) that works! NOT scripts! Q-What things are too hard or annoying to do? Building GUIs, again. Package an image to deploy. Keeping up with the base code. Who need traits, really? Who uses traits, really? It would have been more beneficial to invest that time/effort into cleaning up morphic, modularizing the code, create a GUI builder and an Image Stripper ! Q-What would you like to be able to use Squeak for? Everything. I've always been a Smalltalker! ----------------- Benoît St-Jean Yahoo! Messenger: bstjean Twitter: @BenLeChialeux Pinterest: benoitstjean Instagram: Chef_Benito IRC: lamneth Blogue: endormitoire.wordpress.com "A standpoint is an intellectual horizon of radius zero". (A. Einstein)

11 15

[TEST PLANS] We need your support for making Squeak better
by Marcel Taeumel 26 Jul '18

26 Jul '18

Dear Squeakers, we, the Squeak community, are aiming to do a new release soon. To ensure as high a quality release as possible we'd like to test it adequately. This implies testing across all supported platforms, and possibly on some unsupported ones too. :-) While we have a fairly strong test suite, it does not cover all areas of the system, and not all tests are included such as FFI tests are in its separate package. And of course currently we have about 100 expected failures, 40 unexpected failures, and 10 errors in the base test suite. So we, the board, would like - volunteers to offer to TEST PLATFORMS, different versions of Windows, Raspberry Pi, the many flavors of Linux, etc. Once we have a release candidate we will want to do a clean install on the platform and then run tests. - the community to develop a set of TEST PLANS for testing those parts of the system not tested by the test suite - a REVIEW of the test failures and errors. Are any obsolete? Are any platform-specific? Can we triage them into those that are tractable with little effort, those that are tractable with a lot of effort, those that are insoluble, etc...? Of the expected failures, which are due to fixable bugs, fundamental limitations, etc? So, the data we need to collect includes volunteers, a platform matrix, test plans, and reviews of existing failures and errors. *** To complement Squeak's automated tests, we want to create a MANUAL TEST PLAN. Such a manual plan can be used by any volunteering Squeaker to put any upcoming release "to the acid test." ;-) As a result, our release process gets more robust, and the final release artifact can become of higher quality. As a first step, we want YOU to write down some frequent interactions that you do on a regular basis. You know, things like "I do always do this when adding a class" or "I really need this to do my taxes". It may be a keyboard shortcut, a familiar entry in the world menu, or that lovely button in the upper right corner of tool so-and-so. We want to know! How does it work? Well, we want to guide you a little bit so that things do not get out of hand ... content-wise. ;-) Here are the rules: 1) Give that interaction a name (or title) to express the goal you want to achieve. 2) Briefly describe the starting context. Is the world empty? 3) Divide that interaction into no more than 5 steps so that everybody can click or push his or her way through the Squeak image as quickly as possible. 4) Briefly add some details about the things you should now see or experience. Here is an example: GOAL: "I want to calculate 3+4 in Squeak." CONTEXT: "The world is empty." STEPS - Left click into the world, a menu should appear. - Choose "workspace" from the upper items in the menu. - A workspace window appears, and I click into it's text area. - I type "3+7" and hit [CMD]+[P]. RESULT: "I see a 7." Our idea is to generalize and summarize the most common interactions into a manual test plan for everybody to play around with and report back to the list. When you create a new test, please send it to the list by replying to this email thread. :-) Best, Marcel (for the Squeak Oversight Board)

4 20

[ANN] Changing the Trunk version from 6.0alpha to 5.2alpha
by Marcel Taeumel 24 Jul '18

24 Jul '18

Hi, there. We are going to change the current version number of the Trunk to "5.2alpha" to denote our plans for the upcoming Squeak release. After several discussions within the board and with our release manager Edgar, we came to conclusion that it makes sense to postpone the "6.0" a little bit to finish some bigger concerns such as the use of ephemerons and Etoys clean-up. Best, Marcel

20 55

Bug: SoundService default is not a class
by Nicolas Cellier 19 Jul '18

19 Jul '18

Hi Marcel, Tobias, I encountered a bug, when playing a sound, the default SoundService (class inst var) was not a class (BasedSoundService) but an instance (a BasedSoundService). Why? Because, there is a TileMorphTest>>setUp that remember which is the default, then tries to restore it in tearDown. It used to work. But, in between, SoundService default and defaultOrNil have been re-defined so as to answer an instance rather than a class. Thus, when we play the tests, we can no more play the sounds... I hesitate to open a Mantis issue, no one read nor close them, so the bug report is only here for the moment.

8 21

How does ImageSegmentLoader>>readObject work?
by H. Hirzel 28 Jun '18

28 Jun '18

Hello this is a follow up mail on loading a project file (*.pr) saved in Squeak 3.10.2 [1] into Squeak 6.0a-#18000 [2] The code breaks at ImageSegmentLoader>>readObject [3] I'd like to find out which class causes this problem. readObject | header oop nWords class format | header := self readUint32. (header bitAnd: HeaderTypeMask) caseOf: { [HeaderTypeSizeAndClass] -> [nWords := header >> 2. class := self readUint32. header := self readUint32]. [HeaderTypeClass] -> The header is a 32 bit integer denoting the class, I assume. I'd like to know which class causes the problem. How can I find out which class this 32 bit integer refers to? Regards Hannes -------------------------------------------------------------------------------------------------------------------------------- [1] The project file has a workspace in it in addition to some RectangleMorphs and SimpleButtonMorphs which were loaded succesfully in another test. http://forum.world.st/The-Trunk-Morphic-kfr-1435-mcz-tp5077266p5077476.html The reference has a test project *.pr file attached. [2] http://lists.squeakfoundation.org/pipermail/squeak-dev/2018-May/198848.html New method added SmartRefStream >> in the conversion protocol methodPropertiespps0 ^ AdditionalMethodState [3] readObject | header oop nWords class format | header := self readUint32. (header bitAnd: HeaderTypeMask) caseOf: { [HeaderTypeSizeAndClass] -> [nWords := header >> 2. class := self readUint32. header := self readUint32]. [HeaderTypeClass] -> [class := header - 1. header := self readUint32. nWords := header >> 2 bitAnd: 63]. [HeaderTypeShort] -> [nWords := header >> 2 bitAnd: 63. class := header >> 12 bitAnd: 31]. } otherwise: [self error: 'unexpected free chunk']. nWords := nWords - 1. "nWords includes 1 header word" oop := position. ^[oopMap at: oop ifAbsentPut: [format := header >> 8 bitAnd: 15. "hash := header >> 17 bitAnd: 4095." self allocateObject: format class: class size: nWords]] ensure: [position := oop + (nWords * 4)]

4 5

Dynamic essay project MorphLayoutArticle on Bob's SuperSwiki?
by H. Hirzel 28 Jun '18

28 Jun '18

Hello Is there a backup-copy/mirror available of the MorphLayoutArticle project which was on Bob's SuperSwiki? [1] Regards Hannes ---------------------------------- [1] http://wiki.squeak.org/squeak/2141 How to lay out submorphs Please read the excellent dynamic essay project MorphLayoutArticle (broken link) on Bob's SuperSwiki. Every Morph now has the capability to layout it's submorphs. (Previously, only the AlignmentMorph could implement layout policies. AlignmentMorph is still available because of compatibility reasons and some utility methods it implements.

10 35

Old morphs failing to load in 5.2
by Edgar De Cleene 13 Jun '18

13 Jun '18

http://squeakros.org/5dot2Crashes/Screen%20Shot%202018-05-27%20at%2007.56.4… Here you could see side by side n 3.10 and in Squeak5.2alpha-18019-64bit-All-in-One.app. I wish have some like this in 5.2 so all could see what is new , fixed, etc Squeak style

5 9

Re: [squeak-dev] MorphicProject subclass: #EtoysProject
by LEnglish 08 Jun '18

08 Jun '18

Hmmm? What about if you want to implement a separate hierarchy in a separate System Window, or multiple textures embedded in OpenGL objects? I’m not sure how Craig Latta’s done some of his Caffeine demos, but I doubt if all those Morphs in various independent Canvases are embedded in a single WorldMorph. L "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." -Brian Kernighan > On Oct 3, 2017, at 06:03, Marcel Taeumel <marcel.taeumel(a)hpi.de> wrote: > > +1 :) > > And then later: Rename PasteUpMorph to WorldMorph, and keep an empty PasteUpMorph subclass around for compatibility reasons. So many ideas have been ported down to Morph class over the past years. New applications have no reason to ever use other instances of PasteUpMorph. > > Best, > Marcel >> Am 03.10.2017 14:57:55 schrieb H. Hirzel <hannes.hirzel(a)gmail.com>: >> >> On 10/3/17, H. Hirzel wrote: >> > Dave >> > >> > your change set contains the class EtoysProject with >> > >> > EtoysProject selectors >> > >> > #(#finalEnterActions: #restoreGlobalPreferences #saveGlobalPreferences >> > #initializeProjectPreferences #configureOnFirstEntry >> > #finalExitActions:) >> > >> > For complete configuration of a EtoysProject it might be necessary to do >> > >> > PasteUpMorph subclass: EtoysPasteUpMorph >> > >> > as well. http://wiki.squeak.org/squeak/6461 >> > >> > Then Etoys related methods may be pushed down to EtoysPasteUpMorph. >> >> See screen shot attached. >> >> > And probably an Etoys specific subclass of WorldMenu would be fine as well >> > http://wiki.squeak.org/squeak/6461 >> > >> > >> > there is a test project [2] and some more information about adaptions >> > needed because of the UI changes in the thread 'Etoys in 2017?' - UI >> > preferences [3]. And it would be good to have Etoys methods / >> > configuration separate [4]. >> > >> > I suggest that you start go ahead and start implementing this while >> > using a test Etoys project dropped onto the desktop. >> > >> > --Hannes >> > >> > >> > [2] > You simply drop it in. E.g. download this project >> >> http://etoys.laptop.org/src/Content/ExampleEtoys/CarAndPen.014.pr >> > >> > [3] Hannes Hirzel, 'Etoys in 2017?' mail, Wed, Feb 22, 2017 at 11:01 AM >> > >> > [4] David T. Lewis, Sep 4, 2016 at 3:34 PM >> > "I think it would be great if both Etoys and Scratch were easily >> > loadable and unloadable in trunk." >> > >> > On 10/2/17, David T. Lewis wrote: >> >> An EtoysProject is a project that is configured for running Etoys. On >> >> first entry to a new EtoysProject, the playground and project preferences >> >> are initialized to provide an environment similar to that of a >> >> traditional >> >> standalone Etoys image. >> >> >> >> Certain preferences that are required for Etoys are initialized on >> >> project >> >> entry, overriding their global preference values while this EtoysProject >> >> is active. On leaving the project, these preferences are restored to >> >> their >> >> previous values. >> >> >> >> "ProjectViewMorph openOn: EtoysProject new" >> >> >> >> Change set attached for a minimal implementation. >> >> >> >> Anyone with Etoys knowledge care to help? I do not know enough about >> >> Etoys >> >> to fill in the rest of the initialization that will be required, but it >> >> should not be hard to do. >> >> >> >> Dave >> >> >> >> >> > >> >

3 3

The Trunk: Compiler-eem.381.mcz
by commits＠source.squeak.org 01 Jun '18

01 Jun '18

Eliot Miranda uploaded a new version of Compiler to project The Trunk: http://source.squeak.org/trunk/Compiler-eem.381.mcz ==================== Summary ==================== Name: Compiler-eem.381 Author: eem Time: 31 May 2018, 4:59:26.014435 pm UUID: 6612bdf4-e29a-4cc5-a550-139de9e3c42d Ancestors: Compiler-eem.380 Fix typedScanTokens: for underscore-prefixed symbols such as #_WIN32. =============== Diff against Compiler-eem.380 =============== Item was changed: ----- Method: Scanner>>initScannerForTokenization (in category 'initialize-release') ----- initScannerForTokenization "Use a version of typeTable that doesn't raise xIllegal when enocuntering an _" | underscoreIndex | underscoreIndex := typeTable identityIndexOf: #xUnderscore ifAbsent: [^self]. typeTable := typeTable copy. + typeTable at: underscoreIndex put: #xUnderscoreForTokenization. + typeTable at: (typeTable identityIndexOf: #xLitQuote) put: #xLitQuoteForTokenization! - typeTable at: underscoreIndex put: #xUnderscoreForTokenization! Item was added: + ----- Method: Scanner>>xLitQuoteForTokenization (in category 'multi-character scans') ----- + xLitQuoteForTokenization + "Parse underscores as per the regular parser when following a # symbol, so that e.g. #_WIN32 is correctly tokenized." + | index | + index := typeTable identityIndexOf: #xUnderscoreForTokenization. + typeTable at: index put: #xUnderscore. + [self xLitQuote] ensure: + [typeTable at: index put: #xUnderscoreForTokenization]! Item was changed: ----- Method: Scanner>>xUnderscoreForTokenization (in category 'multi-character scans') ----- xUnderscoreForTokenization + "Parse underscores as assignments so as to be able to correctly tokenize ancient source code." self step. tokenType := #leftArrow. ^token := #'_'!

1 0

Jump to page: