January 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Graphics with Anti-Aliasing

by Guillermo Polito

I cc'ed Carla. I'm sure she should know something about this :) On Sat, Jan 22, 2011 at 11:37 AM, Udo Schneider < udo.schneider(a)homeaddress.de> wrote: > All, > > does somebody has a pointer to a graphics lib/framework/engine for Pharo > which works with Anti-Aliasing. > > My first tests with Balloon2D were disappointing (maybe I'm doing something > wrong). But even setting aaLevel does not seem to bring any result. I.e. the > line drawing code falls back to superclass code which simply draws the same > jagged lines like always. > > Any pointers? > > CU, > > Udo > > >

Jan. 23, 2011

Re: [Pharo-project] default styler is shout

by Guillermo Polito

Anyways, I'm looking that PluggableTextMorph>>textMorphClass "Answer the class used to create the receiver's textMorph" ^TextMorphForEditView and TextMorphForEditView>>editorClass ^ SmalltalkEditor So it's what I suggested needs a lot of work more... On Sat, Jan 22, 2011 at 10:58 PM, Guillermo Polito < guillermopolito(a)gmail.com> wrote: > Ufa,... > > Editor>>styler > ^NullTextStyler new > > SmalltalkEditor>>styler > ^... > > Guille > > > On Sat, Jan 22, 2011 at 10:54 PM, Guillermo Polito < > guillermopolito(a)gmail.com> wrote: > >> Why not delegate it in the used Editor ? SmalltalkEditor should use shout >> while the other editors should use the NullTextStyler. Maybe something like >> this: >> >> PluggableTextMorph>>useDefaultStyler >> self styler: textMorph editor styler >> >> Or better: >> >> PluggableTextMorph>>useDefaultStyler >> self styler: textMorph styler >> >> And then >> >> Editor>>styler >> >> >> >> On Sat, Jan 22, 2011 at 12:25 PM, Tudor Girba <tudor.girba(a)gmail.com>wrote: >> >>> Hi, >>> >>> In the latest Morphic version, the default styler of the >>> PluggableTextMorph is given by: >>> useDefaultStyler >>> "This should be changed to a proper registry but as long as there >>> is only shout this will do" >>> >>> Smalltalk globals >>> at: #SHTextStylerST80 >>> ifPresent: [ :stylerClass | self styler: (stylerClass new >>> view: self) ] >>> ifAbsent: [ self styler: ( NullTextStyler new view: self) >>> ] >>> >>> Thus, if Shout is loaded, you will get it in all text morphs by default. >>> Given that TextMorph should be general and be used in all sorts of contexts, >>> the default behavior should be the original: >>> >>> useDefaultStyler >>> "This should be changed to a proper registry but as long as there >>> is only shout this will do" >>> >>> self styler: (NullTextStyler new view: self; yourself). >>> >>> >>> What do you say? >>> >>> Cheers, >>> Doru >>> >>> -- >>> www.tudorgirba.com >>> >>> "We are all great at making mistakes." >>> >>> >>> >>> >>> >>> >>> >>> >>> >> >

Jan. 23, 2011

Re: [Pharo-project] default styler is shout

by Guillermo Polito

Ufa,... Editor>>styler ^NullTextStyler new SmalltalkEditor>>styler ^... Guille On Sat, Jan 22, 2011 at 10:54 PM, Guillermo Polito < guillermopolito(a)gmail.com> wrote: > Why not delegate it in the used Editor ? SmalltalkEditor should use shout > while the other editors should use the NullTextStyler. Maybe something like > this: > > PluggableTextMorph>>useDefaultStyler > self styler: textMorph editor styler > > Or better: > > PluggableTextMorph>>useDefaultStyler > self styler: textMorph styler > > And then > > Editor>>styler > > > > On Sat, Jan 22, 2011 at 12:25 PM, Tudor Girba <tudor.girba(a)gmail.com>wrote: > >> Hi, >> >> In the latest Morphic version, the default styler of the >> PluggableTextMorph is given by: >> useDefaultStyler >> "This should be changed to a proper registry but as long as there >> is only shout this will do" >> >> Smalltalk globals >> at: #SHTextStylerST80 >> ifPresent: [ :stylerClass | self styler: (stylerClass new >> view: self) ] >> ifAbsent: [ self styler: ( NullTextStyler new view: self) ] >> >> Thus, if Shout is loaded, you will get it in all text morphs by default. >> Given that TextMorph should be general and be used in all sorts of contexts, >> the default behavior should be the original: >> >> useDefaultStyler >> "This should be changed to a proper registry but as long as there >> is only shout this will do" >> >> self styler: (NullTextStyler new view: self; yourself). >> >> >> What do you say? >> >> Cheers, >> Doru >> >> -- >> www.tudorgirba.com >> >> "We are all great at making mistakes." >> >> >> >> >> >> >> >> >> >

Jan. 23, 2011

Re: [Pharo-project] default styler is shout

by Guillermo Polito

Why not delegate it in the used Editor ? SmalltalkEditor should use shout while the other editors should use the NullTextStyler. Maybe something like this: PluggableTextMorph>>useDefaultStyler self styler: textMorph editor styler Or better: PluggableTextMorph>>useDefaultStyler self styler: textMorph styler And then Editor>>styler On Sat, Jan 22, 2011 at 12:25 PM, Tudor Girba <tudor.girba(a)gmail.com> wrote: > Hi, > > In the latest Morphic version, the default styler of the PluggableTextMorph > is given by: > useDefaultStyler > "This should be changed to a proper registry but as long as there is > only shout this will do" > > Smalltalk globals > at: #SHTextStylerST80 > ifPresent: [ :stylerClass | self styler: (stylerClass new > view: self) ] > ifAbsent: [ self styler: ( NullTextStyler new view: self) ] > > Thus, if Shout is loaded, you will get it in all text morphs by default. > Given that TextMorph should be general and be used in all sorts of contexts, > the default behavior should be the original: > > useDefaultStyler > "This should be changed to a proper registry but as long as there is > only shout this will do" > > self styler: (NullTextStyler new view: self; yourself). > > > What do you say? > > Cheers, > Doru > > -- > www.tudorgirba.com > > "We are all great at making mistakes." > > > > > > > > >

Jan. 23, 2011

Re: [Pharo-project] TODO 1.2

by Guillermo Polito

>From the stacktrace, the build is broken because MetacelloVersionNumberTestCase>>testVersion10 is sending a sortBy: message, which is deprecated in pharo. Dale, how can that be fixed? :) On Fri, Jan 21, 2011 at 4:26 AM, Marcus Denker <marcus.denker(a)inria.fr>wrote: > > 1.2 Core: > Bugs: > http://code.google.com/p/pharo/issues/list?can=2&q=milestone=1.2 > Tests: > https://pharo-ic.lille.inria.fr/hudson/job/Pharo%20Core%201.2/lastCompleted… > > 1.2 Full: > Bugs: > http://code.google.com/p/pharo/issues/list?can=2&q=milestone=1.2-DevImage > Tests: N/A, as the buld is broken: > https://pharo-ic.lille.inria.fr/hudson/job/Pharo%201.2/75/console > > There will be *no release of 1.2* if those things are not fixed. If you are > interested, in some way or another, > to have a 1.2 release, you will get it if you help fixing the last bugs. > (hint, hint). > > Marcus > > -- > Marcus Denker -- http://www.marcusdenker.de > INRIA Lille -- Nord Europe. Team RMoD. > > >

Jan. 23, 2011

Re: [Pharo-project] [squeak-dev] Xtreams up to date

by mkobetic＠gmail.com

"Hannes Hirzel"<hannes.hirzel(a)gmail.com> wrote: > If I explore the resulting object I got what is shown in the screen > shot. I am not sure how to interpret the result. Yes, it's not the best possible representation, but it does show the full parse tree. The tree can be rather complicated depending on how complicated is the grammar. That's why I suggested playing with simpler grammars. But the reason why I even mentioned the TreeBuilder is that it's the simplest possible Actor, it implements single method that packages the result of each successful production as it's own node. Maybe using an array instead of associations would create a slightly more "tree like" result. > "Regarding the wiki example: the class PEGWikiGenerator is not included" > > wikiParser parse: 'Page' stream: input actor: PEGWikiGenerator new. > > Where can I get this class? I've attached a fileout of the class that you can file into Squeak, except it references VW XML classes, so it won't work. If you fix it up to use the Squeak XML stuff you should be able to get it to work. All you really need is XML.Element and XML.Tag equivalent. The second fileout shows how the class variables are initialized in VW. You might need to turn the initializers into equivalent class side initialize methods. > As class methods of PEGParser there are 8 grammars included - CSS3, > HttpUrl, JSON, JavaScript, PEG, Smalltalk, Wiki, XML. > > I think a simple complete example which actually does something would > be helpful for me. TreeBuilder seems to do nothing and I do not see the tree. OK, let's try something really simple. The wikipedia page shows a simple arithmetic expression grammar. Here it is transcribed for Xtreams: grammar := ' Number <- [0-9]+ Parenthesised <- "(" Expr ")" Value <- Number / Parenthesised Product <- Value (("*" / "/") Value)* Sum <- Product (("+" / "-") Product)* Expr <- Sum'. I expanded the Value from the wiki grammar into separate Number and Parethesised expression productions. It'll hopefully be clear why shortly. You already know how to get a parser from grammar. That's always the same and for anything more complex you'll probably want to cache the parser instead of recreating it every time, but this will do fine for now: parser := PEG.Parser parserPEG parse: 'Grammar' stream: grammar actor: PEG.ParserParser new. With this you can already parse expressions (note that the grammar as it is written cannot handle whitespace) parser parse: 'Expr' stream: '3+4*(10-4)' reading actor: nil you'll get this (roughly formatted to provide at least some resemblance of the expression tree): #( #(OrderedCollection ($3 "16r0033") OrderedCollection ()) OrderedCollection (# ('+' #(OrderedCollection ($4 "16r0034") OrderedCollection (# ('*' #( '(' #( #(OrderedCollection ($1 "16r0031" $0 "16r0030") OrderedCollection ()) OrderedCollection (# ('-' #(OrderedCollection ($4 "16r0034") OrderedCollection ())))) ')'))))))) That's obviously not very easy to take apart to work with, that's why you want an Actor that will receive rule specific callbacks and do whatever. Let's say we want to evaluate the expression and return the result. We'll make a subclass of Actor called Evaluator. As I mentioned earlier, the Actor class provides pragma based mechanism for expressing rule specific callbacks so you simply write a method for each rule for which you want to do something and annotate it with a pragma. Let's first try to turn the digit collections into numbers: Evaluator>>Number: digits <action: 'Number'> ^digits inject: 0 into: [ :total :digit | total * 10 + ('0123456789' indexOf: digit) - 1 ] With this you can parse numbers parser parse: 'Expr' stream: '3456' reading actor: Evaluator new Now let's try to add. We'll want a method for the Sum production but there's a slightly tricky part there. Note that the rule is defined as Sum <- Product (("+" / "-") Product)* which means that you get a term (Product) and then zero or more pairs of + or - and another term. What the method receives will be collection of the first term and then pairs of the operator with its term. So parsing this: parser parse: 'Expr' stream: '3+4+5+6+7' reading actor: Evaluator new will give you #(3 OrderedCollection (#('+' 4) #('+' 5) #('+' 6) #('+' 7))) So if you add a method for Sum, that's what it will need to deal with: Sum: terms <action: 'Sum'> ^terms last inject: terms first into: [ :total :next | (next first = '+') ifTrue: [ total + next last ] ifFalse: [ total - next last ] ] The Actor pragmas provide another convenience for taking more complex input apart. You can write the above as Sum: first rest: pairs <action: 'Sum' arguments: #(1 2)> ^pairs inject: first into: [ :total :pair | (pair first = '+') ifTrue: [ total + pair last ] ifFalse: [ total - pair last ] ] The arguments: keyword simply says invoke this method with elements 1 and 2 of the rule input (terms) as arguments. Obviously the method must take the corresponding number of arguments as well. So with the above rules the addition expression should give you 25. The Product rule handling is analogous: Product: first rest: pairs <action: 'Product' arguments: #(1 2)> ^pairs inject: first into: [ :total :pair | (pair first = '*') ifTrue: [ total * pair last ] ifFalse: [ total / pair last ] ] Finally the parenthesised expression is rather simple: Parenthesised: expression <action: 'Parenthesised' arguments: #(2)> ^expression It receives a three element array: the left bracket, the expression inside and the right bracket. All we really need to do is return the expression in the middle. The above is using the arguments: pragma again to achieve that, the following would be equivalent: Parenthesised: expression <action: 'Parenthesised'> ^expression at: 2 I hope by now it's also clearer why it was useful to factor the Value definition out into separate Number and Parenthesised rule. Otherwise we'd have to handle both cases in a single Value: method, which would make the code messier. > For example besides the classes PEGParserPEGTest and > PEGParserSmalltalkTest (6 test cases) a class PEGParserHttpUrlTest or > PEGParserJavaScriptTest would be helpful. Yup, the tests are seriously wanting as well. Hopefully we'll get there soon. BTW, the parsing stuff doesn't work for me out of the box in either Squeak or Pharo in the images that I use (which should be quite clean). The problem I'm seeing is that the Dictionary>>addAll: in ParserParser>>Grammar: doesn't work as expected. It's trying to add collection of associations to a dictionary but ends up adding them as values indexed by integers. I can fix it but I wonder what do people have loaded that makes it work for them. Anyway, hopefully the above can get you going. Cheers, Martin

Jan. 23, 2011

[Pharo-project] closeable tabs

by Tudor Girba

Hi, I worked a bit on adding the possibility of closing tabs from a tab group. This is available for the moment in the I refactored the LazyTabGroupMorph from the Glamour repository. It would still be good to delegate a couple of things to the Theme, but other than that I think we could replace the logic from TabGroupMorph. What I did was to introduce a LazyTabPage that can lazily draw the page morph. If you (Alain, Gary?) have time, I could use a code review :). Gofer new squeaksource: 'Glamour'; package: 'ConfigurationOfGlamour'; load. (Smalltalk at: #ConfigurationOfGlamour) perform: #loadDefault Cheers, Doru -- www.tudorgirba.com "In a world where everything is moving ever faster, one might have better chances to win by moving slower."

Jan. 23, 2011

Re: [Pharo-project] IMO IllegalResumeAttempt should be resumable

by Eliot Miranda

Belay that. Apologies for the noise. resumeUnchecked: is of course the right solution. e.g. Array streamContents: [:s| [Cogit chooseCogitClass testPCMappingSelect: [:m| true]] on: AssertionFailure, Error do: [:ex| | ctxt | ctxt := ex signalerContext findContextSuchThat: [:ctx| ctx sender selector = #allSelect:]. "testPCMappingSelect uses allSelect: to enumerate so return to allSelect: to continue" s nextPut: (ctxt tempAt: 1). Transcript ensureCr; cr; print: (ctxt tempAt: 1); cr; flush. ex searchFrom: ctxt. ex resumeUnchecked: false]] On Sat, Jan 22, 2011 at 1:50 PM, Eliot Miranda <eliot.miranda(a)gmail.com>wrote: > Hi All, > > an attempt to resume an unresumable signal results in > raising IllegalResumeAttempt; quite right: > > Exception>>resume > "Return from the message that signaled the receiver." > > self resume: self defaultResumeValue > > > However, IllegalResumeAttempt is not itself resumable. So in those rare > cases (e.g. test coverage) where a knowledgeable client wants to force > resumption of a non-resumable signal by catching and resuming > IllegalResumeAttempt, they can't because > > IllegalResumeAttempt>>isResumable > ^ false > > > So take e.g. > > [Cogit chooseCogitClass testPCMappingSelect: [:m| true]] > on: AssertionFailure > do: [:ex| | ctxt | > ctxt := ex signalerContext findContextSuchThat: > [:ctx| ctx sender selector = #allSelect:]. > Transcript ensureCr; print: (ctxt tempAt: 1); flush. > [ex resume] > on: IllegalResumeAttempt > do: [:ex| ex resume]] > > where I'm trying to JIT every method in my image and catalogue those that > assert-fail, proceeding through the assert failure (and AssertionFailure > isn't resumable). I happen to know that proceeding from the assert-fails is > safe. But I can't force resumption because IllegalResumeAttempt is itself > not resumable. > > I propose that we make IllegalResumeAttempt resumable. Objections? > Cautions? > > best > Eliot >

Jan. 22, 2011

Re: [Pharo-project] Some important news

by GermÃ¡n Arduino

+1 2011/1/22 Esteban Lorenzano <estebanlm(a)gmail.com>: > +1 > > El 21/01/2011, a las 11:47p.m., Tony Fleig escribiÃ³: > >> Excellent. Where do I sign up? >> >> On Fri, Jan 21, 2011 at 11:26 AM, StÃ©phane Ducasse >> <stephane.ducasse(a)inria.fr> wrote: >>> Hi guys >>> >>> some important news today, we decided that we will set up a pharo association >>> so that we can promote pharo. We would like to be able to collect money and >>> create bounties to improve pharo. >>> >>> Stef >>> >> > > > -- ================================================= GermÃ¡n S. ArduinoÂ <gsa @ arsol.net>Â Â Twitter: garduino Arduino Software & Web HostingÂ Â http://www.arduinosoftware.com PasswordsProÂ http://www.passwordspro.com =================================================

Jan. 22, 2011

[Pharo-project] IMO IllegalResumeAttempt should be resumable

by Eliot Miranda

Hi All, an attempt to resume an unresumable signal results in raising IllegalResumeAttempt; quite right: Exception>>resume "Return from the message that signaled the receiver." self resume: self defaultResumeValue However, IllegalResumeAttempt is not itself resumable. So in those rare cases (e.g. test coverage) where a knowledgeable client wants to force resumption of a non-resumable signal by catching and resuming IllegalResumeAttempt, they can't because IllegalResumeAttempt>>isResumable ^ false So take e.g. [Cogit chooseCogitClass testPCMappingSelect: [:m| true]] on: AssertionFailure do: [:ex| | ctxt | ctxt := ex signalerContext findContextSuchThat: [:ctx| ctx sender selector = #allSelect:]. Transcript ensureCr; print: (ctxt tempAt: 1); flush. [ex resume] on: IllegalResumeAttempt do: [:ex| ex resume]] where I'm trying to JIT every method in my image and catalogue those that assert-fail, proceeding through the assert failure (and AssertionFailure isn't resumable). I happen to know that proceeding from the assert-fails is safe. But I can't force resumption because IllegalResumeAttempt is itself not resumable. I propose that we make IllegalResumeAttempt resumable. Objections? Cautions? best Eliot

Jan. 22, 2011