- Pharo-dev - lists.pharo.org

Re: [Pharo-project] What happened to COTDC?

by Sean P. DeNigris

We can't afford to take a break! (SystemNavigation default allClassesAndTraits reject: [:aClass| aClass hasComment]) size "1131" -- View this message in context: http://forum.world.st/What-happened-to-COTDC-tp4110480p4113481.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Nov. 27, 2011

Re: [Pharo-project] Morphic replacement

by Sean P. DeNigris

Fernando olivero-2 wrote > > found much more productive the approach of rewriting it from scratch. > Me too. I always started using Morphic, but every time I got stuck, I would go deeper and deeper and always end up starting to make my own TextMorph, lol. Fernando olivero-2 wrote > > No World as a subclass of PasteUpMorph > Glad to hear that. PasteUpMorph has identity issues :) Fernando olivero-2 wrote > > Hope that very soon i can provide more details, but my plans are to > release it as part of the next Gaucho version, to provide a working > example. > Excellent, thanks! -- View this message in context: http://forum.world.st/Morphic-replacement-tp4106870p4113468.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Nov. 27, 2011

Re: [Pharo-project] What happened to COTDC?

by Sean P. DeNigris

StÃ©phane Ducasse wrote > > but somebody else can do it. > Okay, I'll start again tomorrow. -- View this message in context: http://forum.world.st/What-happened-to-COTDC-tp4110480p4113451.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Nov. 27, 2011

Re: [Pharo-project] Building 32-bit Cairo library on Macs

by Fernando Olivero

nice! thanks a lot for dealing with the ugliness outside the smalltalk world, so we mortals can deal with higher level languages. Looking forward to see an Athens binding to Cairo. Fernando On Fri, Nov 25, 2011 at 7:33 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > cool > We want more :) > > Stef > > On Nov 25, 2011, at 5:17 PM, Igor Stasenko wrote: > >> Just to let you know: >> >> it works on Mac! >> I just loaded the code which Javier did for linux, changed path to >> library and it worked out of the box. >> all examples in [1] are working out of the box! >> >> So, now i can start wiring Cairo backend for Athens :) >> >> [1] http://squeaksource.com/NBCairo/ >> >> -- >> Best regards, >> Igor Stasenko. >> > > >

Nov. 27, 2011

Re: [Pharo-project] Morphic replacement

by Fernando Olivero

When the vm, starts announcing multitouch events, it's a matter of adding that particular peripheral device to the model. For now, the vm only announces mouse and keyboard events, therefore i only modeled those. Thanks for the question, maybe we should ask the vm guys to add this to their todo list. Fernando On Sat, Nov 26, 2011 at 8:15 PM, Douglas Brebner <squeaklists(a)fang.demon.co.uk> wrote: > On 25/11/2011 10:00, Fernando Olivero wrote: > > Sounds good except maybe >> The world has a mouse and a keyboard, which take care of events >> announced by the world. > > Does that mean you're not thinking of support for multitouch? > > >

Nov. 27, 2011

Re: [Pharo-project] Morphic Cleanup Status

by Fernando Olivero

Differently from what i've shown you at ESUG, i didn't reify the events. The vm announces a new event to the GUserInterface, who dispatches the event to the mouse or the keyboard. Which in turn, according to its inner state (modeled using a fsm), sends a message to the user interface, such as #clickedAt: , #secondaryClickedAt: , #move: #moveDragging:, #keystroke: aGKeystroke, etc.. The following diagram depicts the scheme. Fernando On Fri, Nov 25, 2011 at 10:38 PM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: > > On Nov 25, 2011, at 11:08 AM, Fernando Olivero wrote: > >> Hi Sean, i got overwhelmed last year with the task of porting CUIS >> Morphic, to Pharo. There's just to much differences on several common >> classes, it's hard to fully understand how to merge both ui >> frameworks. >> >> Even though i believe Juan has a did an excellent work in simplifying >> Morphic, and adding simpler code, like for example his new layouts >> (download CUIS and try them out). >> >> But my overall impression , was that we need to stop being backward >> compatible in order to create a simple Morphic-like framework for >> Pharo, and start from scratch. Fixing current Morphic is also a >> humongous Â task, for example see Stef's nice work on cleaning the >> event handling mechanism. > > BTW you should be able to use the new event hierarchy: it is clean (no first secondâ¦..) and a nice support for double > dispatch. I hope that next week I can sit with igor and we finish our journey. > How do you deal with event in Gaucho. > > Stef >> >> Fernando >> >> On Fri, Nov 25, 2011 at 9:38 AM, Alain Plantec <alain.plantec(a)yahoo.com> wrote: >>> Le 25/11/2011 07:39, Sean P. DeNigris a Ã©crit : >>>> What's currently in progress? SimpleMorphic? CUIS related? Other? >>>> >>>> I was pulling my hair out with Morphic layouts yesterday and started to >>>> clean, but I don't want to duplicate efforts on things already being worked >>>> on. There are a bunch of open issues and discussions on the list. What's >>>> actively being done right now? >>> Hi Sean, >>> AFAIK, no deep cleaning action. nothing from my side for now. >>> I think they have done some refactoring around layout managing in CUIS >>> (SimpleMorphic). >>> btw, **thanks a lot** for all you are doing. >>> Cheers >>> Alain >>> >>>> >>>> Thanks. >>>> Sean >>>> >>>> -- >>>> View this message in context: http://forum.world.st/Morphic-Cleanup-Status-tp4106566p4106566.html >>>> Sent from the Pharo Smalltalk mailing list archive at Nabble.com. >>>> >>> >>> >> > >

Nov. 27, 2011

Re: [Pharo-project] [squeak-dev] [ANN] First try: Fuel for Squeak

by David T. Lewis

On Sun, Nov 27, 2011 at 03:34:27PM -0300, Mariano Martinez Peck wrote: > > > > Hi Mariano, > > > > I loaded Fuel into my working Squeak image, which is an older image kept up > > to date with trunk via the update stream (i.e. it is "bleeding edge" with > > various additional packages loaded). > > > > > Excellent :) > Where can I get a "bleeding edge" Squeak image? The most recent update on the ftp server is <ftp://ftp.squeak.org/4.3alpha/Squeak4.3gamma-11793.zip> Run that image, then from the world menu (left click anywhere), select help... -> update code from server When the update is complete, you have a fully updated image. > > > I had to implement SystemVersion>>major and SystemVersion>>minor to > > complete > > the installation (these are Pharo idioms, so I just implemented them to > > answer 0 on Squeak). > > > > > To complete the instalation?? Can you remember why (the sender) it is > needing #mayor and #minor? > As far as I remember, it is only used by Fuel in FLBasicSerializationTest > >> expectedFailures > but that is not the installation, but running tests. I'm not sure, but for whatever reason it popped up as an error while I was installing. > > > > I did not encounter the timeout problem, but if you have a long running > > test, just add a pragma like <timeout: 30> to set the timeout to 30 > > seconds. > > This will increase the timeout when running on Squeak, and should not > > affect Pharo. > > > > Thanks, I didn't know. Is there a way to run the tests in the test runner > so that it doesn't take into account such timeouts? > or can I change the global preference? I ask so that to avoid changing > tests code (even if they would work for Pharo). I don't think that there is a preference for this. It's probably best to just provide the pragma for the specific tests, since this will not affect the tests themselves. > Thanks for the detailed answer. From what I can see, even if those tests > are wrong, "most cases" should work fine. Yes, it is working fine for me. I just changed RemoteTask to make use of Fuel for serializing objects between images, and it seems to work very well. Dave

Nov. 27, 2011

Re: [Pharo-project] [squeak-dev] [ANN] First try: Fuel for Squeak

by Mariano Martinez Peck

On Sun, Nov 27, 2011 at 3:25 PM, David T. Lewis <lewis(a)mail.msen.com> wrote: > On Sat, Nov 26, 2011 at 05:06:17PM -0300, Mariano Martinez Peck wrote: > > Hi guys. After a couple of efforts from Martin Dias and I, we have a > > "working" Fuel serializer for Squeak. I used the quotes because all I > could > > do is to run tests. > > Fuel is a binary serializer and you can find all the information in > > http://rmod.lille.inria.fr/web/pier/software/Fuel > > This is a project sponsored by ESUG SummerTalk and mainly developed by > > Martin Dias. > > Fuel seems to work fine in Squeak (despite some issues in the tests), so > I updated RemoteTask <http://wiki.squeak.org:8080/squeak/6176> to use Fuel > rather than ReferenceStream for returning the results of a remote task to > the supervisor image. If Fuel is not present in the image, ReferenceStream > will be used by default. This update is included in the latest CommandShell > on SqueakSource. > > Excellent. I am glad to hear that. Let us know if there are problems. > Fuel provides provides a noticable performance improvement when returning > a large result data set, and also avoids a problem in serializing large > integers <http://bugs.squeak.org/view.php?id=7679> . > Thanks. Indeed, the performance is one of the most important characteristic of Fuel. > > Pharo users should beware of this bug when running the RemoteTask examples > <https://code.google.com/p/pharo/issues/detail?id=4997>, because the > examples perform prime number searches in the remote tasks. > Ok, it seems the issue is being discussed. Cheers > > Dave > > > -- Mariano http://marianopeck.wordpress.com

Nov. 27, 2011

Re: [Pharo-project] [squeak-dev] [ANN] First try: Fuel for Squeak

by Mariano Martinez Peck

> > Hi Mariano, > > I loaded Fuel into my working Squeak image, which is an older image kept up > to date with trunk via the update stream (i.e. it is "bleeding edge" with > various additional packages loaded). > > Excellent :) Where can I get a "bleeding edge" Squeak image? > I had to implement SystemVersion>>major and SystemVersion>>minor to > complete > the installation (these are Pharo idioms, so I just implemented them to > answer 0 on Squeak). > To complete the instalation?? Can you remember why (the sender) it is needing #mayor and #minor? As far as I remember, it is only used by Fuel in FLBasicSerializationTest >> expectedFailures but that is not the installation, but running tests. > > I did not encounter the timeout problem, but if you have a long running > test, just add a pragma like <timeout: 30> to set the timeout to 30 > seconds. > This will increase the timeout when running on Squeak, and should not > affect Pharo. > Thanks, I didn't know. Is there a way to run the tests in the test runner so that it doesn't take into account such timeouts? or can I change the global preference? I ask so that to avoid changing tests code (even if they would work for Pharo). > > Below are some notes that I took while running the tests. The most > important > issue would be the tests that fail in CompiledMethod comparison. I cannot > say if this is a Fuel issue or just something about the way the tests are > written, but I note significant differences in the implementation of > CompiledMethod>>= between Squeak and Pharo, so it is definitely something > worth checking. > > That's a good point. There must be something there for sure. > Notes on test issues in FuelTests: > > FLSequencedSerializationTest>>assertSequencedSerializationEqualityOf: > is failing due to difference in CompiledMethod>>= between Pharo and > Squeak. Not sure if this is a Fuel issue. > > FLBasicSerializationTest>>testGradientFillStyle fails because > GradientFillStyle>>= > is an identity test, whereas in Pharo it has been added from package > '*PolyMorph-Widgets' (i.e. not part of base Pharo either). > > FLBasicSerializationTest>>testMethodDictionary fails due to compiled > method #= > (same issue as > FLSequencedSerializationTest>>assertSequencedSerializationEqualityOf:) > > In FLBlockClosureSerializationTest, all tests fail with an assertion > failure > in MethodContext > > FLFullBasicSerializationTest>>testMethodDictionaries fails due to the > compiled method equality check problem. > > FLMethodContextSerializationTest tests fail with assertion failure in > MethodContext > > FLSortedCollectionSerializationTest tests fail with assertion failure > in MethodContext > > FLCreateClassSerializationTest has errors due to DNU from > ClassForTestToBeDeleted1 class>>basicLocalSelectors > > > FLCompiledMethodSerializationTest>>testTwoCompiledMethodsReferencingSameClassVariable > fails due to the compiled method equality check problem. > > FLCreateTraitSerializationTest>>testCreateBasic fails with DNU > on Trait>>basicLocalSelectors > > > Thanks for the detailed answer. From what I can see, even if those tests are wrong, "most cases" should work fine. Cheers -- Mariano http://marianopeck.wordpress.com

Nov. 27, 2011

Re: [Pharo-project] [squeak-dev] [ANN] First try: Fuel for Squeak

by David T. Lewis

On Sat, Nov 26, 2011 at 05:06:17PM -0300, Mariano Martinez Peck wrote: > Hi guys. After a couple of efforts from Martin Dias and I, we have a > "working" Fuel serializer for Squeak. I used the quotes because all I could > do is to run tests. > Fuel is a binary serializer and you can find all the information in > http://rmod.lille.inria.fr/web/pier/software/Fuel > This is a project sponsored by ESUG SummerTalk and mainly developed by > Martin Dias. Fuel seems to work fine in Squeak (despite some issues in the tests), so I updated RemoteTask <http://wiki.squeak.org:8080/squeak/6176> to use Fuel rather than ReferenceStream for returning the results of a remote task to the supervisor image. If Fuel is not present in the image, ReferenceStream will be used by default. This update is included in the latest CommandShell on SqueakSource. Fuel provides provides a noticable performance improvement when returning a large result data set, and also avoids a problem in serializing large integers <http://bugs.squeak.org/view.php?id=7679> . Pharo users should beware of this bug when running the RemoteTask examples <https://code.google.com/p/pharo/issues/detail?id=4997>, because the examples perform prime number searches in the remote tasks. Dave

Nov. 27, 2011