March 2013 - Pharo-dev - lists.pharo.org

[Pharo-project] new version of the (old) versionner
by Alexandre Bergel March 31, 2013

March 31, 2013

Hi! In the meantime we have a reliable solution to manage configuration, I have updated http://squeaksource.com/Versionner.html The two most significant changes are: - SmalltalkHub is now supported - Committing a project now produces a stable version (instead of a development) and the most recent baseline is marked as #development. In the meantime to have Versionner2 around, I worked a bit on Versionner to solve some pressing issues with Moose. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

2 1

[Pharo-project] Request write access to PharoInbox
by Nicolas Cellier March 31, 2013

March 31, 2013

Hello, can someone give me write access to PharoInbox? Nicolas Cellier

3 6

[Pharo-project] GSoC deadline passed, record number of ideas collected
by Janko MivÅ¡ek March 31, 2013

March 31, 2013

Dear Smalltalkers, Serge and me just submitted the application for the GSoC (see its copy here: http://gsoc2013.esug.org/application) now let we wait finger crossed to the Monday 8.April when accepted organizations will be announced. But we have a good chance thanks to all of you spend a time to prepare the record number of ideas: 44, which is way over last year 25! Note also that you can still propose an project idea, there is two more weeks or so for that, but we need a solid list of ideas at that time for Google guys to see what we are really capable :) While idea authors please recheck that your idea is properly put on http://gsoc2013.esug.org/ideas. For any mistake I apologize in advance! Thanks therefore again to all, now a week and so a rest! Best regards Serge and Janko your GSoC admin team -- Janko MivÅ¡ek Smalltalk GSoC Admin Team http://gsoc2013.esug.org

3 2

Re: [Pharo-project] PDF Rendering
by Christian Haider March 31, 2013

March 31, 2013

My name is Haider (sticky fingers are not polite with proper names...). My PDF library has nothing to do with rendering (putting the graphics on the screen) and I doubt that Artefact will do that. IMHO, rendering PDF is quite a challenge... Cheers, Christian (HAIDER) > -----UrsprÃ¼ngliche Nachricht----- > Von: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project- > bounces(a)lists.gforge.inria.fr] Im Auftrag von stephane ducasse > Gesendet: Sonntag, 31. MÃ¤rz 2013 11:16 > An: Pharo-project(a)lists.gforge.inria.fr > Betreff: Re: [Pharo-project] PDF Rendering > > Else there is also the work of Christian Haidler. > Artefact developers will release soon a 0.9 version once they fix the jpeg > inclusion Stef > > >> I was looking around for PDF rendering libraries that might be used with > NativeBoost and they all seem to be licensed GPL, which is not compatible > with Pharo's preference for the MIT license. Then I came across Apache > licensed pdf.js [1] which sparked my curiosity.... There are a few projects to > have Smalltalk run on top of Javascript, but has anyone looked at the > reverse, running javascript inside a Smalltalk image? Perhaps handing the > Javascript off via NativeBoost to a Javascript VM like SquirrelFish [2]. Or is > that just a bad idea? > >> > >> cheers -ben > >> > >> [1] http://andreasgal.com/2011/06/15/pdf-js/ > >> [2] https://www.webkit.org/blog/189/announcing-squirrelfish/ > > > > Maybe you already knew about Artefact - a PDF framework for Pharo > > > > http://forum.world.st/Artefact-a-PDF-framework-td4585877.html > > > > A solution in Smalltalk is more useful in the long term, IMHO, in the sense > that it can then be read, understood, maintained by the community instead > of being a black box. > > > > Sven > > > > -- > > Sven Van Caekenberghe > > http://stfx.eu > > Smalltalk is the Red Pill > > > > > > >

2 1

[Pharo-project] TxText progress and discussion
by Denis Kudriashov March 31, 2013

March 31, 2013

Hello. I implement positioning text for new TxTextMorph. You can now see blinking cursor, move it right, left, up and down, position it by mouse click. I make new version 0.4 which restore jenkins builds https://ci.inria.fr/rmod/job/TxText/, You can try TxTextMorph class side example methods. Details: Cursor position at "visual text" presented by TxTextLayoutPosition. TxTextLayoutPosition has state: - textLayout - span - TxLayoutSpan (introduced by Igor suggestion) - spanPosition - index inside TxLayoutSpan - desiredOffset - desired x offset from start of "visual line". It can differ from actual (x,y) position - xyValue - actual (x,y) position inside "visual text" box. TxTextLayoutPosition implements moveRight. moveLeft, moveUp and moveDown methods. Desired offset allow moving up and down with remember of first offset. Example: 123456 123 123456 Cursor moves from up to bottom. Position of cursor at each line will be after red char. It is common behaviour of text editors. But our current text morph not works such way. And current text morph implements move up and move down in terms of index from start line. So it is visually incorrect up or down movement for no mono fonts. Try this to check it: WWWWWW - 6 W-characters, move from 6-th W-char down and try same at pharo. 1234567890 TxLayoutSpan present part of text span which placed at some line and has offset, extent and font. It was created by layout builder to fit text layout restrictions. When text span wrapped by max line width restriction It "splitted" for two or more LayoutSpan's. TxLayoutSpan has span interval (TxInterval). Igor was suggest removing explicit font attribute from text model and replace it with small font parts like family name, bold, underlined and others. He said it's should be responsibillity of "view system" to select appropriate font by this attributes. I don't know is it good idea. But I have some user cases where it is needed. So TxLayoutSpan allows such design decision. That's why it has font instance. TxLayoutSpan is created by layout builder which can select appropriate font for underlying spans. Now it just extracts font from span attributes like before. There is TxCursorMorph which presents blinking cursor (basic morph stepping logic). So visual text cursor is just submorph of TextMorph. I not yet implement any communication level between TxTextLayout, TxTextLayoutPosition and morphs. It is for future. Now morphs implemented to show what already works. I needs help with namings. All ideas are welcome. For example we have too much different position names: - TxPosition with vars linePosition, spanPosition - indexes - TxTextLayoutPosition with vars spanPosition (index of char), desiredOffset (x coordinate of offset inside line), xyValue (x@ycoordinates inside text box) - TxCharactersSpan methods: #isLastPosition: , #lastInsertPosition - indexes - TxLayoutSpan methods: #charPositionAtOffset: (char index at offset). #computeXYPositionAt:, #containsXPosition:, #endPosition (TxPosition). I want easily understand what kind of position method means. In TxLayoutSpan I try to distinct it by "charPosition" and "xyPosition". I dont sure it is nice. And I don't like name of TxLayoutSpan. (too many spans). Small tasks which I have in mind: - more cursor movement actions - make settings for shortcuts at TxTextMorph - make cursor actions dynamically connected to shortcut settings. Each cursor action (method) should automatically creates shortcut setting for it. There is strange problem with TxTextMorph when it is transparent. Try change #drawOn: method to see it (remove rectangle filling). Next I will implement events logic to update TxLayoutPosition when layout changed. And text selection stuff. Best regards, Denis

3 4

[Pharo-project] FileSystem-Legacy fixes
by Johan Brichau March 31, 2013

March 31, 2013

Hi Pharoers, Hi Camillo, To make Seaside3.0.x work in Pharo2.0, we included the FileSystem-Legacy package. (this is probably the easiest solution for Seaside3.0.x branch) I did need to make some changes because this package had a slightly different behavior than the legacy code in some methods: - add #primPathNameDelimiter - #directoryExists: should not only check if it exists, it should also check if it's a directory: (self resolve: filenameOrPath) exists and:[(self resolve: filename) isDirectory] - #directoryNamed: was actually able to accept a fully qualified pathName as an argument (not logical, but that's legacy ;-). So implemented as follows: FileDirectory on: (self fullNameFor: localFileName) What do you think? Can I publish these changes to the FileSystem package? regards Johan

2 2

[Pharo-project] a glimpse at new cool text editor made by alain P ; )
by StÃ©phane Ducasse March 31, 2013

March 31, 2013

Thanks alain for this previea and Synectique ( http://www.synectique.eu ) for the sponsoring of this effort. Stef

7 6

[Pharo-project] starting a debugger on a caught error
by Ben Coman March 31, 2013

March 31, 2013

I am trying to debug a red-square-of-death for a Morph. I have found where Morph errors are handled in Morph>>fullDrawOn: where it goes... [ "stuff to draw morphs" ] on: Error do: [ :err | "stuff to draw red square" ]. In the do: block I'd like to open a debugger on the error. I see Exception has #debug, so I changed the block to... [ :err | err inspect. "stuff to draw red square" ] ...and then in the inspector I do (self debug) ...but this brings up the emergency evaluator. Below is a contrived example that exhibits the same behaviour. doIt on... (1 / 0. self inform: 'x'.) and of course a debugger appears. doIt on... [ 1 / 0. self inform: 'x' . ] on: Error do: [:err | err inspect ] ...and then (self debug) in the inspector you get the emergency evaluator. Anyone know how can I get the debugger to come up on 'err' in the last example? cheers -ben

2 2

[Pharo-project] obtaining a FileUrl through String>>asUrl
by Tudor Girba March 30, 2013

March 30, 2013

Hi, I am trying to get to a FileUrl via the polymorphic asUrl call, but it seems I am doing something wrong. I do: 'file:d:\' asUrl but the problem seems that $: and $\ are mangled, because when I print, I get: file:///d%3A%5C Furthermore: 'file:d:\' asUrl asFileReference exists ==> false but, at the same time: 'd:\' asFileReference exists ==> true Can anyone tell me where is the mistake? Btw, the reason I am looking for this is that I would like to use the ZeroConf support to load a configuration from a directory Monticello repository and to do that I have to pass a url to the command line. Cheers, Doru -- www.tudorgirba.com "Every thing has its own flow"

4 7

[Pharo-project] [update 3.0] #30018
by Marcus Denker March 30, 2013

March 30, 2013

1 0