- Pharo-dev - lists.pharo.org

[Pharo-project] [ANN] 10330
by StÃ©phane Ducasse June 8, 2009

June 8, 2009

- Issue 867: PluggableButtonMorph>>getModelState - Issue 868: MouseOverHandler problem stef

1 0

[Pharo-project] Fixes I had to make to the latest pharo image to be able to work with it
by Hernan Wilkinson June 8, 2009

June 8, 2009

Hi, I know this is not the right way to send fixes, but it is the fastest... so maybe it is useful for those who need them. I had a lot of problems using the last pharo image, some are related with the new debugger, some are old errors related with buttons, some are related with the OmmiBrowser (I send them here too because I know that the OB authors are subscribed to this list and because that is the main pharo browser), etc. Most of the problems are because nil receives a message that does not understand. I did not look for the reason that made the variable reference nil (maybe the best solution) but only added a #isNil check (the fastest solution). I had to disable the new debugger and get back to the old one (Debugger) because I was getting to many errors with the new one saying "It could not debug a running process". Here are the fixes: 1) This one made the image to crash because it entered a infinite loop. The problem is that sometime getStateSelector is nil. PluggableButtonMorph>>getModelState "Answer the result of sending the receiver's model the getStateSelector message." ^(getStateSelector isNil or: [ model isNil ]) ifTrue: [false] ifFalse: [model perform: getStateSelector] --------------- 2) This one is related with what Sean Allen reported. Again leftMorphs is nil, so I added a #isNil check. (For me this class is a mess, but I do not have the experience to change it yet) MouseOverHandler>>noticeMouseOver: aMorph event: anEvent "Remember that the mouse is currently over some morph" leftMorphs isNil ifFalse: [ (leftMorphs includes: aMorph) ifTrue:[leftMorphs remove: aMorph] ifFalse:[enteredMorphs nextPut: aMorph]. overMorphs nextPut: aMorph.] -------------- 3) OCompletion does not work with the old debugger because#guessTypeForName: references tempNames that is not defined... So I just disallowed it. Debugger>>guessTypeForName: aString | index object | index := nil. "tempNames indexOf: aString ifAbsent: []." object := index ifNil: [index := self receiver class allInstVarNames indexOf: aString ifAbsent: []. index ifNil: [^ nil]. self receiver instVarAt: index] ifNotNil: [self selectedContext tempAt: index]. ^ object class -------------- All the rest are OB related. 4) When pressing the mouse right button (on windows) I had a error saying OBTextSelection does not understand #isClassNode, so I implemented returning false. Again, I did not verify if an OBTextSelection should receive that message or not, I just implemented. OBTextSelection>>isClassNode ^false --------------- 5) After loading a Monticello package I could not open an OB browser anymore. Again, the problem is that ea is nil sometimes. OBBrowser>>scanNodeCommands: ann (cmdFactories select: [:ea | ea notNil and: [ ea takesNodes]]) do: [:ea | ann addFactory: ea] ------------------- 6) Similar to 5. OBBrowser>>scanTextCommands: ann (cmdFactories select: [:ea | ea notNil and: [ ea takesText ]]) do: [:ea | ann addFactory: ea] ------------------ 7) This is an error I reported to OB a long time ago, related with pressing or selecting the cancel option. OBDefinitionPanel>>aboutToChange: ann | ans answerResult vetoBlock | self canDiscardEdits ifTrue: [^ true]. vetoBlock := [ ann veto ]. ans := OBChoiceRequest prompt: 'Code has been modified. Do you want to accept or discard the changes?' labels: {'Accept'. 'Discard'. 'Cancel'} values: (Array with: [self changed: #accept] with: [self changed: #clearUserEdits] with: vetoBlock ). answerResult := ans value. answerResult isNil ifTrue: vetoBlock. -------------------- I hope it helps. Hernan.

3 4

[Pharo-project] [10329] removed TabbedPalette
by StÃ©phane Ducasse June 8, 2009

June 8, 2009

Thanks alain :) Stef

1 0

[Pharo-project] ObjectFinder
by StÃ©phane Ducasse June 8, 2009

June 8, 2009

Hi guys Yesterday I got a nice demo of ObjectFinder builds on top of OB. It is a really good improvement over the chasing browser. I really suggest that you get a look at it. http://decomp.ulb.ac.be/frdricpluquet/personalstuff/ stef

7 16

[Pharo-project] Win32Setup for 10324dev09.06.1
by Torsten Bergmann June 8, 2009

June 8, 2009

Within the pharo files page [1] you will now also find a new Win32 Setup [2] based on the latest "pharo0.1-10324dev09.06.1.". This now also includes the fonts as provided by Damien. Attached you will find the new modified NSIS script - based on [3]. Janko wrote >So, there is at least some short text needed to encourage user to >explore further, at least a reference to Squeak by example book. And how >to open a Class browser ... I expect to soon have a "Pharo by example" from Stephane and friends?! ;) If so I will package the PDF right into the setup... Depending on time we can also finish the "Getting started workspace" Have fun Torsten [1] https://gforge.inria.fr/frs/?group_id=1299 [2] https://gforge.inria.fr/frs/download.php/22410/setup_pharo_0.0.0.3.exe [3] http://lists.gforge.inria.fr/pipermail/pharo-project/2009-May/008790.html [4] http://code.google.com/p/pharo/wiki/GettingStartedWorkspaces -------------------------------------------------------------------------- -- GRATIS fÃ¼r alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

2 1

[Pharo-project] Debugger problem with pragmas
by Nicolas Cellier June 8, 2009

June 8, 2009

I don't know if it's only me, but I sometimes can't open the debugger... I've got a problem with this code (eem): MethodProperties>>analogousCodeTo: aMethodProperties pragmas ifNil: [aMethodProperties pragmas notEmpty ifTrue: [^false]] ifNotNil: [aMethodProperties pragmas empty ifTrue: [^false]. snip... What is this #empty message? Shouldn't it be isEmpty or ifEmpty: [] ? If I correct it, I have further problems because #analogousCodeTo: is not implemented in Pragma Nicolas

1 0

[Pharo-project] Call for suggestions: Morph has 997 methods
by StÃ©phane Ducasse June 7, 2009

June 7, 2009

Hi guys It would be great if we could assess Morphic fat. I think that with 997 methods not counting the MorphExtension class, we should be able to clean Morph and friend. So if you see passing unused, useless, broken Morph code please report it Stef

2 2

[Pharo-project] an odd bug
by Sean Allen June 7, 2009

June 7, 2009

that i dont really know how to describe... so i'm going to attach a screenshot... i opened the lastest pharo dev image... checked to see if the redraw bug that i have been experiencing was still there... went to quick and got a sudden flurry of message not understood errors ( see screenshot of the rearranged windows )... i wasnt able to reliably reproduce but, when doing the same thing ( open and quit... ) about 2 out of every five times i got message not understood windows popping up. [image: message not understood.png]

8 21

[Pharo-project] Citezen loading (Rio-Kernel defect)
by Schwab,Wilhelm K June 7, 2009

June 7, 2009

Damien, I found the problem: #newForWin32 is missing the #new, so it sends the class to do an instance's job. I fixed that, but Monitcello was determined to undo my efforts. After a few tries at getting around that, I realized that it should run on Linux, so I did the installation there and then added my fix so that it should run on win32 also. That appears to have worked. I find that it cannot parse my "real" .bib files, but it does seem to handle things that I build up from entries from Google scholar. The bib files in question are tested in the sense the LaTeX/BibTeX are happy with them, but I have yet to try bibclean on them. Is there any documentation or example code available? Bill

4 4

[Pharo-project] can we remove TabbedPalette
by Alain Plantec June 7, 2009

June 7, 2009

Hi all, I know it is used by the Smacc GUI. I can provide a new Smacc GUI implementation (old one is really messy) any other good reason to keep TabbedPalette ? thanks Alain

3 5