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

Re: [Pharo-dev] Making ASTInterpreter work with opal bindings

by Camillo Bruni

Sorry I wasn't very well yesterday so I skipped a couple of mails... If you want we can pair-programm some time on that (and a possible paper?) On 2013-05-15, at 13:57, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > Clement? Camilo? :D > On Tue, May 14, 2013 at 2:46 PM, Guillermo Polito <guillermopolito(a)gmail.com >> wrote: >> On Tue, May 14, 2013 at 2:44 PM, Guillermo Polito < >> guillermopolito(a)gmail.com> wrote: >> >>> >>> On Tue, May 14, 2013 at 1:03 PM, ClÃ©ment Bera <bera.clement(a)gmail.com>wrote: >>> >>>> Hey Guille, >>>> >>>> To debug AST-interpreter there is a specific AIContextInspector. What >>>> you can do is just to inspect in the debugger the latest AIContext created >>>> by the AST-interpreter and you should see the stack of AIContext in the >>>> inspector. It is not perfect but it helps a lot. There is also the >>>> AST-Debugger but it is just a prototype so you can hardly use it. >>>> >>> >>> Buu, how do I get the ASTDebugger? Maybe I can give it a try... >>> >>> >>>> >>>> testEnsureTricky is the one that checks ensure nested in the on:do: >>>> works correctly. Basically it checks when you unwind the context up to the >>>> exception handler that you executes the ensure blocks code. It's kind of >>>> the trickiest one to fix :/. In the exception chapter of Deep into Pharo I >>>> added a section somewhere in Exception implementation about that, it might >>>> help you to read it. >>>> >>>> Take care, AST-interpreter tests might not be enough. You should check >>>> the console output of the AST-interpreter-regression build too. (here : >>>> https://ci.inria.fr/rmod/job/ast-interpreter-regression/89/console) But >>>> it is failing since some new updates in Pharo 3.0 ... >>>> >>> >>> I'm looking at https://ci.inria.fr/rmod/job/ast-interpreter-regression/. >>> What does this job test/do? We can add that to the description :) >>> >>> Anyway, I'm loading the same as the job, so I can test the slice locally >>> in my machine :) >>> >> >> And I just did: >> >> 68 tests run >> 68 tests were successful >> 0 tests failed >> 0 tests were too long to be run >> >> Sound like good results, but I have nothing to reflect with :). >> >> >>> >>> >>>> >>>> Are you committing on the AST-interpreter repository or on Pharo >>>> directly ? >>>> >>> >>> Pharo... Should I commit to the other repo? Don't have grants =^). >>> >>> >>>> I don't think there are but there might be some fixes already on the dev >>>> branch > Smalltalkhub user: dh83 project: ast-interpreter >>>> >>> >>> Will check. Actually, I just loaded bleeding edge + merged my slice, and >>> the same test still fails. >>> >>> >>>> >>>> I would like to work on Opal test this afternoon but if you come to >>>> lille this week (not today because there is the strike) we can have a look >>>> together. >>>> >>> >>> Tomorrow! >>> >>> >>>> For M4te I've also fixed some stuff in Hz on the m4te repo (basically >>>> compatibility with the new RB visitor API) it might be nice to commit it on >>>> your repo too. >>>> >>> >>> cool :) >>> >>> >>>> >>>> See you, >>>> >>>> >>>> >>>> 2013/5/14 Guillermo Polito <guillermopolito(a)gmail.com> >>>> >>>>> Right now, in the image there exist two ways to annotate an AST: >>>>> - the Opal way >>>>> - the RB Way >>>>> >>>>> We should simplify that by having only one. So far, the only user of >>>>> the RB bindings is the ASTInterpreter + the RB tests. >>>>> >>>>> So, I prepared a slice and issue [1], with the following details >>>>> >>>>> - annotate ast using opal annotator >>>>> - extended OCBindings to read and write from an AST interpreter >>>>> - not yet made a difference between arg nodes with a different class >>>>> (just put an if) >>>>> - i didnt add inst vars into the opal bindings for the binding->node >>>>> relation as in RB, I send the node as a parameter >>>>> - I had to adapt the #compile: into ASTInterpreter because >>>>> - opal does not annotate other than method nodes >>>>> - undeclared variables raised warnings, not errors >>>>> >>>>> There is one problem yet, I have only one failing test: >>>>> #testEnsureTricky. I don't understand yet why it is failing and debugging >>>>> it is pretty complex :). So any help is welcome in reviewing the changes. >>>>> >>>>> Guille >>>>> >>>>> [1] >>>>> https://pharo.fogbugz.com/f/cases/10610/Make-ASTInterpreter-use-OC-Opal-bin… >>>>> >>>> >>>> >>>> >>>> -- >>>> ClÃ©ment BÃ©ra >>>> Mate Virtual Machine Engineer >>>> BÃ¢timent B 40, avenue Halley 59650 *Villeneuve d'Ascq* >>>> >>> >>> >>

May 15, 2013

AsmJit refactoring

by Camillo Bruni

I started to work on a basic ARM model for AsmJit with Damien Pollet and on the fly we continue the refactoring I started last week: - no more PoolDictionaries - no more global registers - reducing the number of class variables - more specific register classes for x86 to avoid nasty bit masking/checking operations on an integer All in all this makes the code much more accessible and you can be sure that changing a register (aka renaming it) is a local cation. What do you think?

May 15, 2013

Re: [Pharo-dev] Making ASTInterpreter work with opal bindings

by Marcus Denker

On May 15, 2013, at 1:58 PM, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > Clement? Camilo? :D > I will have a look with Clement this afternoon. > > On Tue, May 14, 2013 at 2:46 PM, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > > > > On Tue, May 14, 2013 at 2:44 PM, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > > On Tue, May 14, 2013 at 1:03 PM, ClÃ©ment Bera <bera.clement(a)gmail.com> wrote: > Hey Guille, > > To debug AST-interpreter there is a specific AIContextInspector. What you can do is just to inspect in the debugger the latest AIContext created by the AST-interpreter and you should see the stack of AIContext in the inspector. It is not perfect but it helps a lot. There is also the AST-Debugger but it is just a prototype so you can hardly use it. > > Buu, how do I get the ASTDebugger? Maybe I can give it a try... > > > testEnsureTricky is the one that checks ensure nested in the on:do: works correctly. Basically it checks when you unwind the context up to the exception handler that you executes the ensure blocks code. It's kind of the trickiest one to fix :/. In the exception chapter of Deep into Pharo I added a section somewhere in Exception implementation about that, it might help you to read it. > > Take care, AST-interpreter tests might not be enough. You should check the console output of the AST-interpreter-regression build too. (here : https://ci.inria.fr/rmod/job/ast-interpreter-regression/89/console) But it is failing since some new updates in Pharo 3.0 ... > > I'm looking at https://ci.inria.fr/rmod/job/ast-interpreter-regression/. What does this job test/do? We can add that to the description :) > > Anyway, I'm loading the same as the job, so I can test the slice locally in my machine :) > > And I just did: > > 68 tests run > 68 tests were successful > 0 tests failed > 0 tests were too long to be run > > Sound like good results, but I have nothing to reflect with :). > > > > Are you committing on the AST-interpreter repository or on Pharo directly ? > > Pharo... Should I commit to the other repo? Don't have grants =^). > > I don't think there are but there might be some fixes already on the dev branch > Smalltalkhub user: dh83 project: ast-interpreter > > Will check. Actually, I just loaded bleeding edge + merged my slice, and the same test still fails. > > > I would like to work on Opal test this afternoon but if you come to lille this week (not today because there is the strike) we can have a look together. > > Tomorrow! > > For M4te I've also fixed some stuff in Hz on the m4te repo (basically compatibility with the new RB visitor API) it might be nice to commit it on your repo too. > > cool :) > > > See you, > > > > 2013/5/14 Guillermo Polito <guillermopolito(a)gmail.com> > Right now, in the image there exist two ways to annotate an AST: > - the Opal way > - the RB Way > > We should simplify that by having only one. So far, the only user of the RB bindings is the ASTInterpreter + the RB tests. > > So, I prepared a slice and issue [1], with the following details > > - annotate ast using opal annotator > - extended OCBindings to read and write from an AST interpreter > - not yet made a difference between arg nodes with a different class (just put an if) > - i didnt add inst vars into the opal bindings for the binding->node relation as in RB, I send the node as a parameter > - I had to adapt the #compile: into ASTInterpreter because > - opal does not annotate other than method nodes > - undeclared variables raised warnings, not errors > > There is one problem yet, I have only one failing test: #testEnsureTricky. I don't understand yet why it is failing and debugging it is pretty complex :). So any help is welcome in reviewing the changes. > > Guille > > [1] https://pharo.fogbugz.com/f/cases/10610/Make-ASTInterpreter-use-OC-Opal-bin… > > > > -- > ClÃ©ment BÃ©ra > Mate Virtual Machine Engineer > BÃ¢timent B 40, avenue Halley 59650 Villeneuve d'Ascq > > >

May 15, 2013

Re: [Pharo-dev] Making ASTInterpreter work with opal bindings

by Guillermo Polito

Clement? Camilo? :D On Tue, May 14, 2013 at 2:46 PM, Guillermo Polito <guillermopolito(a)gmail.com > wrote: > > > > On Tue, May 14, 2013 at 2:44 PM, Guillermo Polito < > guillermopolito(a)gmail.com> wrote: > >> >> On Tue, May 14, 2013 at 1:03 PM, ClÃ©ment Bera <bera.clement(a)gmail.com>wrote: >> >>> Hey Guille, >>> >>> To debug AST-interpreter there is a specific AIContextInspector. What >>> you can do is just to inspect in the debugger the latest AIContext created >>> by the AST-interpreter and you should see the stack of AIContext in the >>> inspector. It is not perfect but it helps a lot. There is also the >>> AST-Debugger but it is just a prototype so you can hardly use it. >>> >> >> Buu, how do I get the ASTDebugger? Maybe I can give it a try... >> >> >>> >>> testEnsureTricky is the one that checks ensure nested in the on:do: >>> works correctly. Basically it checks when you unwind the context up to the >>> exception handler that you executes the ensure blocks code. It's kind of >>> the trickiest one to fix :/. In the exception chapter of Deep into Pharo I >>> added a section somewhere in Exception implementation about that, it might >>> help you to read it. >>> >>> Take care, AST-interpreter tests might not be enough. You should check >>> the console output of the AST-interpreter-regression build too. (here : >>> https://ci.inria.fr/rmod/job/ast-interpreter-regression/89/console) But >>> it is failing since some new updates in Pharo 3.0 ... >>> >> >> I'm looking at https://ci.inria.fr/rmod/job/ast-interpreter-regression/. >> What does this job test/do? We can add that to the description :) >> >> Anyway, I'm loading the same as the job, so I can test the slice locally >> in my machine :) >> > > And I just did: > > 68 tests run > 68 tests were successful > 0 tests failed > 0 tests were too long to be run > > Sound like good results, but I have nothing to reflect with :). > > >> >> >>> >>> Are you committing on the AST-interpreter repository or on Pharo >>> directly ? >>> >> >> Pharo... Should I commit to the other repo? Don't have grants =^). >> >> >>> I don't think there are but there might be some fixes already on the dev >>> branch > Smalltalkhub user: dh83 project: ast-interpreter >>> >> >> Will check. Actually, I just loaded bleeding edge + merged my slice, and >> the same test still fails. >> >> >>> >>> I would like to work on Opal test this afternoon but if you come to >>> lille this week (not today because there is the strike) we can have a look >>> together. >>> >> >> Tomorrow! >> >> >>> For M4te I've also fixed some stuff in Hz on the m4te repo (basically >>> compatibility with the new RB visitor API) it might be nice to commit it on >>> your repo too. >>> >> >> cool :) >> >> >>> >>> See you, >>> >>> >>> >>> 2013/5/14 Guillermo Polito <guillermopolito(a)gmail.com> >>> >>>> Right now, in the image there exist two ways to annotate an AST: >>>> - the Opal way >>>> - the RB Way >>>> >>>> We should simplify that by having only one. So far, the only user of >>>> the RB bindings is the ASTInterpreter + the RB tests. >>>> >>>> So, I prepared a slice and issue [1], with the following details >>>> >>>> - annotate ast using opal annotator >>>> - extended OCBindings to read and write from an AST interpreter >>>> - not yet made a difference between arg nodes with a different class >>>> (just put an if) >>>> - i didnt add inst vars into the opal bindings for the binding->node >>>> relation as in RB, I send the node as a parameter >>>> - I had to adapt the #compile: into ASTInterpreter because >>>> - opal does not annotate other than method nodes >>>> - undeclared variables raised warnings, not errors >>>> >>>> There is one problem yet, I have only one failing test: >>>> #testEnsureTricky. I don't understand yet why it is failing and debugging >>>> it is pretty complex :). So any help is welcome in reviewing the changes. >>>> >>>> Guille >>>> >>>> [1] >>>> https://pharo.fogbugz.com/f/cases/10610/Make-ASTInterpreter-use-OC-Opal-bin… >>>> >>> >>> >>> >>> -- >>> ClÃ©ment BÃ©ra >>> Mate Virtual Machine Engineer >>> BÃ¢timent B 40, avenue Halley 59650 *Villeneuve d'Ascq* >>> >> >> >

May 15, 2013

Re: [Pharo-dev] Doubt about RING

by Igor Stasenko

On 15 May 2013 11:29, Sebastian Tleye <stleye(a)gmail.com> wrote: > Hello > > I am making some modifications in the smalltalk meta-model and i need > to replace classes with Ring classes (it is supposed to be possible since > they are polymorphic) > But i am having a little problem, all the classes know how to answer the > message "isObsolete" > but Ring classes doesn't, to solve this i have defined the method > "isObsolete" in RGBehaviorDefinition > but i am not sure which code does it need > > - if i copy and paste the code from Behavior, then i obtain more > "doesNotUnderstand" messages > since the code from Behavior>>isObsolete is "self instanceCount = 0" but > Ring classes doesn't know how to answer "instanceCount" neither > so i suspect this is not the right way to do it. > > - if i copy the code from Class, then it seems it works, but the result from > for example > "Array isObsolete" is different of the result from "Array asRingDefinition > isObsolete" consequently well.. #isObsolete for model classes doesn't makes any sense. i would just answer false to this message. and i would ask why this message is ever sent. because this is not something essential to class model. > next messages obtain different results and finally a a "doesNotUnderstand" > message again. > > Am i doing something wrong? Any suggestion? -- Best regards, Igor Stasenko.

May 15, 2013

Refactoring API and the AST Nodes

by Gisela Decuzzi

I've been thinking about this... because the refactoring api can be a pain to handle (at least for me) in some cases, specially when I'm working with an AST node... because in the node it's all (or almost all) the information needed for the api and I started to ask me... why I'm always opening the node to extract all the strings just to use the Refactorings? I don't like to do that, and I think that time to time we have that problem and ended repeating code... I was thinking in change the refactoring API to accept an AST node, but it's not true that we have the node in the hand all the time we want to use it, and adding more entry points could complex the api. Also if we convert a node in a model for the refactors there are some cases where the information isn't complete in the node, by example to rename a variable you need the new name... and that is not present in the node but needed for the refactor... so is not that simple to make them an inocent model... Then I thought that I can use a new object that takes the node, and creates the refactors needed (similar to NautilusRefactoring using an ast node), but in this model I think that we are doing yet another work around and not a real solution... The other day Guille points me that maybe I should think that the refactoring API is kind of ok... and maybe the nodes could give me the refactors or an almost configured refactor... since they have (almost all) the information. So... I was thinking in this approach, it could be a solution to all the strings handling, and it kind of make sense that the node knows the refactors that apllies and how to construct it with the information needed. A number of ideas... and I would like to discuss about them, maybe here is someone who has the same problem and arrived to a different solution, so I would like to exchange opinions and improve that part that I think we are working around but not solving :(

May 15, 2013

Re: [Pharo-dev] [Pharo-users] success story: NTed

by Manfred KrÃ¶hnert

Hi Norbert, I have two remarks to your many valuable input. On Tue, May 14, 2013 at 5:37 PM, Norbert Hartl <norbert(a)hartl.name> wrote: > Hi Nico, > > > I do have one question: what was your experience with Amber? > > > > as an overall statement I would say it was a pleasure to use amber to > develop the app. The initial setup until the class browser opens is easy. > Learning how to create packages and knowing that code persistence is based > on packages is also quite ok because amber includes a WebDav Server that > can be started from the downloaded amber repo. So no big problems in > starting. > > Thinks that go less smooth: > > - the web app I'm developing is checked into git. Amber puts all files (js > and st) in the same directory. This makes managing files in version control > system a bit cumbersome. > - all code is put in the same directory. The amber code checked out from > the repository is in the same place as the code that I produced myself. > Again you need to do additional steps to do to clean it up > - tools for combining code are missing. That is not a big problem because > a simple cat on the console can solve this but would be nice not have to. > The Amber commandline compiler can do that partially. Maybe we can figure out together what is still missing to get it working as expected. > - I prepared two different locations on my web server for the same app. > One that is used to develop (the WebDav server access it) and one where > only the combined javascript source code is used. This is needed to test > the target deployment. Something that could ease this would be wonderful > > What keeps me from using it in every project: > > - The startup times for amber can be quite huge. All the class > initialization is a lot of stuff that burns CPU cycle. On an iphone you > don't notice really but on the Nexus 7 tablet it takes 2-4 seconds for > initialization and the first page to appear. This is a not a problem for > desktop browsers but for mobile ones where I find amber extremely good for. > I only used amber in pre pre 1.0 versions and I know there have been > optimizations that I didn't test, yet. > > And the annoying "You didn't ask for but I add it anyway wishlist": > > - I need to test Helios because I think it solves a lot of issues with the > current browser. I don't know how it is supposed to work that you have a > second window with the IDE that accesses a page. But I'm sure it works > flawlessly or I misunderstood it from the screenshots > - retarget files on the disc. I would like to have different paths on disc > for separate packages. > - amber using websockets to connect to a backend pharo server. This can be > used for comfortable code management, amber support for state transport > client->server and server->client and to replicate console log to the pharo > transcript (or whatever). This is important on mobile devices where > sometimes you cannot access the browser console. > - it would be fine to have support for obfuscation. It is not only for > hiding code but also for making it smaller thus saving bandwidth and > startup times. This could be the case if the produced javascript code would > be work with google closure compiler. Even better would be to have visitor > for AST/IR to make it even more effective/less dangerous. > Here again, the Amber commandline compiler is capable of using the Google closure compiler to minify/obfuscate Amber programs. However, I only tested this completely with server side applications. Here again, it would be nice if we could resolve outstanding or unknown issues. > > Norbert > Best, Manfre > > Am 14.05.2013 um 15:32 schrieb Nicolas Petton <petton.nicolas(a)gmail.com>: > > > Nico > > > > > > On May 14, 2013, at 11:37 AM, Norbert Hartl <norbert(a)hartl.name> wrote: > > > >> I think my last post of this message fell into the time the old mailing > list dies. So here it is again > >> > >> --- > >> > >> We (@2denker in cooperation with a local university) built a monitoring > tool for disaster scenario recovery rehearsals. Several times a year > universities, red cross, THW und hospitals are planning a rehearsal > together. For this an environment is prepared for a simulated disaster (car > accidents, explosions in civil areas, â¦). The local environment is prepared > to look real as well as film make-up artist prepare people for injuries > (broken legs, open head, â¦). > >> The rehearsal should improve the coordination of several hospitals, > firemenâ¦as well as the quickest possible treatment for the injured. > >> > >> One important part of the rehearsal is monitoring of what happens. This > is achieved by installing a local position system and using our tool. Every > injured person gets a wireless tablet. On the tablet there are UI elements > that indicate treatment done to them. Whenever a doctor comes visiting the > injured person activates the corresponding UI button. The events are > transmitted to a server and are prepared for another analysis software. The > tablet enables the HQ to make contact to individual persons or groups. The > tablet measures times of needed treatments and instructs the injured to be > dead if measurements weren't taken in time. > >> > >> Technically it is done in smalltalk solo. > >> > >> The server is a pharo image. It has > >> > >> - web admin interface with complex forms for managing injuries, > injured, groups, etc. For this we used seaside because it is best suited > for this job. Forms are done with magritte, magritte-bootstrap. The > interface also displays the live incoming events and other data > >> - a static file server that delivers static resources to client tablets > >> - a websocket connector that is used by the clients to connect to the > server > >> - a http handler to export data to other analysis tools > >> > >> The client is done with amber smalltalk. It > >> > >> - loads the web page containing amber from the server > >> - connects back to the server > >> - downloads a UI spec for the injured using the tablet to display > personalized UI elements > >> - stores events on the tablet in case the injured is offline > >> - measures times to instruct the injured > >> > >> If you have any questions please don't hesitate to ask. > >> > >> Norbert > > > > -- > > Nicolas Petton > > http://www.nicolas-petton.fr > > > > > > >

May 15, 2013

Doubt about RING

by Sebastian Tleye

Hello I am making some modifications in the smalltalk meta-model and i need to replace classes with Ring classes (it is supposed to be possible since they are polymorphic) But i am having a little problem, all the classes know how to answer the message "isObsolete" but Ring classes doesn't, to solve this i have defined the method "isObsolete" in RGBehaviorDefinition but i am not sure which code does it need - if i copy and paste the code from Behavior, then i obtain more "doesNotUnderstand" messages since the code from Behavior>>isObsolete is "self instanceCount = 0" but Ring classes doesn't know how to answer "instanceCount" neither so i suspect this is not the right way to do it. - if i copy the code from Class, then it seems it works, but the result from for example "Array isObsolete" is different of the result from "Array asRingDefinition isObsolete" consequently next messages obtain different results and finally a a "doesNotUnderstand" message again. Am i doing something wrong? Any suggestion?

May 15, 2013

[update 3.0] #30131

by Marcus Denker

[update 3.0] #30131 30131 ----- 10596 Set class for #compiler in Behavior>>#compiler https://pharo.fogbugz.com/f/cases/10596 10627 add workaround for failing ZNTests https://pharo.fogbugz.com/f/cases/10627 10630 move FamFam package to Polymorph https://pharo.fogbugz.com/f/cases/10630 Diff information: http://smalltalkhub.com/mc/Pharo/Pharo30/main/Zinc-Tests-MarcusDenker.179.d… http://smalltalkhub.com/mc/Pharo/Pharo30/main/Polymorph-Widgets-MarcusDenke… http://smalltalkhub.com/mc/Pharo/Pharo30/main/Polymorph-TaskbarIcons-Marcus… http://smalltalkhub.com/mc/Pharo/Pharo30/main/FamFam-Icons-MarcusDenker.25.… http://smalltalkhub.com/mc/Pharo/Pharo30/main/Compiler-MarcusDenker.458.diff

May 15, 2013

Re: [Pharo-dev] vm fixes

by Benjamin

Thank you :) I am only using image side primitives, so any Vm should handle it (I am using DisplayScreen hostSize: / hostTitle:) Ben On May 15, 2013, at 10:44 AM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > This is beautiful. I guess this only works on Mac, right? > > Cheers, > Doru > > > On Tue, May 14, 2013 at 10:51 PM, Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> wrote: > In the following picture, instead of doing the classical > MessageBrowser new openWithSpec, and get a Morphic window inside the os window of Pharo > > I do > MessageBrowser openWorldWithSpec > And here the message browser window *is* the system window. > Basically, it sticks the morph on the world, do some magic around, and voilÃ . > The cool part is that the rest of the code do not change at all, and still it opens a the good size, the title is properly updated etc > > <Screen Shot 2013-05-14 at 4.44.00 PM.png> > > Ben > > On May 14, 2013, at 10:28 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: > >> This sounds intriguing. Can you be more specific? :) >> >> Doru >> >> >> On May 14, 2013, at 10:23 PM, Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> wrote: >> >>> Thank you :) >>> >>> This fix goes with pair with >>> https://pharo.fogbugz.com/default.asp?10619 >>> >>> Thanks to Esteban, one can now deploy a Spec widget with the os window as window :) >>> >>> Ben >>> >>> On May 14, 2013, at 10:19 PM, stephane ducasse <stephane.ducasse(a)free.fr> wrote: >>> >>>> thanks esteban! >>>> >>>> On May 14, 2013, at 9:54 PM, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: >>>> >>>>> Hi, >>>>> >>>>> I just pushed fix for a weird bug in cocoa cog vm. >>>>> It was crashing image when resizing using DisplayScreen>>#hostWindowSize: but even worst, some times was crashing also when dragging a window outside the world screen... >>>>> >>>>> The fix is there, but you will need to wait until tomorrow to have it in zeroconf (because of weird mac jenkins slave permissions). >>>>> >>>>> This increases a bit more the VM stability, still not perfect, but we are getting closer :) >>>>> >>>>> Esteban >>>> >>>> >>> >> >> -- >> www.tudorgirba.com >> >> "Presenting is storytelling." >> >> > > > > > -- > www.tudorgirba.com > > "Every thing has its own flow"

May 15, 2013