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

Re: [Pharo-dev] [update 3.0] #30425

by Marcus Denker

On Sep 27, 2013, at 9:25 AM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > And it's interesting that during the first pass on a dirty image it > works well even on the slave too. > I will revert the compiler merge for now (it did even run the regression testsâ¦ but showing some strange - I think - trait-recompiling related bug..) So what we will do -> revert this update (it is already auto-reverted due to the crash) -> save Opal as it is in the image before this to to the Opal repo -> get the Opal regression green (traits bug) -> get image green (we do have some failing tests) -> *then* on the Opal repo, merge step-by-the intermediate 5 small commits (Abstract compiler class, some trivial cleanups, clean block analysis that is not called) I do not see how these changes could lead to the problem, but it could be that there is some bug in the compiler that shows itself when recompiling certain methods. One interesting observation: we did not check the Opal-Regression tester for 4 weeks, and we are in the mess. This just shows how important continous integration + automatic regression testing *after every commit* is for compiler developmentâ¦ (or being more intelligent would maybe another solution ;-) Marcus

Sept. 27, 2013

Re: [Pharo-dev] UI Mess

by Goubier Thierry

Le 27/09/2013 01:14, Camillo Bruni a Ã©crit : > > On 2013-09-26, at 15:50, GOUBIER Thierry <thierry.goubier(a)cea.fr> wrote: > >> Thanks, Sean, I'm happy that after Erwan suggestion of an issue, someone took charge of cleaning that :) >> >> My own position is that current Pharo has unified the filesystem issue with FileReference, and that I'm not impressed by moving ten variants of an API for choosing files from one class to another. But that's just me :) >> >> I think at the core we have a single method which returns FileReferences, and which has four parameters: title, start path, pattern, name (no need for an open). I'll move some of the API complexity in the pattern object, so that it may be able to make the distinction between files and directory, and of course extensions (with a reflective method to list the possible extensions). The pattern should be easy to create :), i.e. nil, '*.png', #isADirectory, whatever. All parameters may be nil, and like that there isn't any dependency on Morphic or the command line. >> >> answer := UIManager default >> chooseFileReference: 'Select a file' >> path: aFileReference >> pattern: '*.png' >> name: nil >> >> (name: is for those Save As dialogs where you already have a pre-existing file name) > > I would also say that spans over the most common options. Convert #pattern: to #filter: > which takes a block over the shown file reference and you can implement almost everything ;) The block idea is good, but I'd prefer something that the GUI can query for some information, such as the list of file patterns (which is often listed in those file open, file save as dialog boxes). Hence the FilePattern idea. Thierry -- Thierry Goubier CEA list Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95

Sept. 27, 2013

Re: [Pharo-dev] UI Mess

by Goubier Thierry

For the theme idea, I allways have in mind the structure used in VisualWorks which is to have decorator wrappers (and a few specifics like scrollbars) for each theme, and the core GUI stays the same. Morphs aren't designed that way (except for scrollbars) so it looks harder. But maybe composing would work (a morph may have a border and delegates some behavior to the border: drawing, bounds, etc... Changing the theme would mean changing the border class. In a few cases, it would mean changing the morph class completely (i.e. scrollbars)). Designing and testing that sounds like long, hard work :( For the UIManager theme thing, I would couple that to Smalltalk, something as simple as : Smalltalk ui (command line or gui) and: Smalltalk ui theme (if someone want to write an ascii-art Morphic theme :)) (I used to write dialog boxes and menu bars in ascii for a VT100 terminal a long, long time ago). Thierry Le 26/09/2013 22:14, StÃ©phane Ducasse a Ã©crit : > If you remember a while ago I was thinking that the theme could be managed like settings (ie encapsulated in the widgets and that the widgets could encapsulate their default for theme), but sadly I was wrong because a theme oftn requires specific computation. So probably > we should have a widget and some object that represents the theme strategy. > Ideally I would like to get rid of the UIManager theme idea but this is difficult and it requires a bit of prototyping. > So if one of you want to try to see on a typical case how a design could work. It would be really great. > > Setf > -- Thierry Goubier CEA list Laboratoire des Fondations des SystÃ¨mes Temps RÃ©el EmbarquÃ©s 91191 Gif sur Yvette Cedex France Phone/Fax: +33 (0) 1 69 08 32 92 / 83 95

Sept. 27, 2013

Re: [Pharo-dev] [regression reporter]regression occurred

by Marcus Denker

It gets strange by the minuteâ¦ Are we sure that the VM did not change? On Sep 27, 2013, at 9:41 AM, no-reply(a)ci.inria.fr wrote: > https://ci.inria.fr/pharo/job/Pharo-3.0-Update-Step-2.1-Validation/label=wi… > > 20 regressions found. > CollectionsTests.Weak.WeakFinalizersTest.testNewFinalizationSupported > CollectionsTests.Weak.WeakRegistryTest.testFinalization > CollectionsTests.Weak.WeakRegistryTest.testFinalizationWithMultipleFinalizersPerObject > CollectionsTests.Weak.WeakRegistryTest.testFinalizationWithOnFork > CollectionsTests.Weak.WeakRegistryTest.testGrowingByAHighPriorityProcessDoesntLeak > CollectionsTests.Weak.WeakRegistryTest.testGrowingDoesntLeak > CollectionsTests.Weak.WeakRegistryTest.testRemovingByAHighPriorityProcessDoesntLeak > CollectionsTests.Weak.WeakRegistryTest.testRemovingDoesntLeak > Graphics.Tests.Files.PNGReadWriterTest.testBlue16 > Graphics.Tests.Files.PNGReadWriterTest.testBlue32 > Graphics.Tests.Files.PNGReadWriterTest.testBlue8 > NativeBoost.Tests.NBCoreTests.testExternalResourceManagerFinalization > NativeBoost.Tests.NBCoreTests.testFinalization > NativeBoost.Tests.NBInterpreterProxyTest.testStackPointer > NativeBoost.Tests.NBInterpreterProxyTest.testStackPointer2 > Tests.Finalization.ObjectFinalizerTests.testFinalizationOfMultipleResources > Tests.Monticello.MCDirectoryRepositoryTest.testAddAndLoad > Tests.Monticello.MCDirectoryRepositoryTest.testIncludesName > Tests.Monticello.MCDirectoryRepositoryTest.testLoadMissingNode > Tests.Monticello.MCDirectoryRepositoryTest.testStoreAndLoad >

Sept. 27, 2013

[regression reporter]regression occurred

by no-reply＠ci.inria.fr

https://ci.inria.fr/pharo/job/Pharo-3.0-Update-Step-2.1-Validation/label=wi… 20 regressions found. CollectionsTests.Weak.WeakFinalizersTest.testNewFinalizationSupported CollectionsTests.Weak.WeakRegistryTest.testFinalization CollectionsTests.Weak.WeakRegistryTest.testFinalizationWithMultipleFinalizersPerObject CollectionsTests.Weak.WeakRegistryTest.testFinalizationWithOnFork CollectionsTests.Weak.WeakRegistryTest.testGrowingByAHighPriorityProcessDoesntLeak CollectionsTests.Weak.WeakRegistryTest.testGrowingDoesntLeak CollectionsTests.Weak.WeakRegistryTest.testRemovingByAHighPriorityProcessDoesntLeak CollectionsTests.Weak.WeakRegistryTest.testRemovingDoesntLeak Graphics.Tests.Files.PNGReadWriterTest.testBlue16 Graphics.Tests.Files.PNGReadWriterTest.testBlue32 Graphics.Tests.Files.PNGReadWriterTest.testBlue8 NativeBoost.Tests.NBCoreTests.testExternalResourceManagerFinalization NativeBoost.Tests.NBCoreTests.testFinalization NativeBoost.Tests.NBInterpreterProxyTest.testStackPointer NativeBoost.Tests.NBInterpreterProxyTest.testStackPointer2 Tests.Finalization.ObjectFinalizerTests.testFinalizationOfMultipleResources Tests.Monticello.MCDirectoryRepositoryTest.testAddAndLoad Tests.Monticello.MCDirectoryRepositoryTest.testIncludesName Tests.Monticello.MCDirectoryRepositoryTest.testLoadMissingNode Tests.Monticello.MCDirectoryRepositoryTest.testStoreAndLoad

Sept. 27, 2013

Re: [Pharo-dev] [update 3.0] #30425

by Pavel Krivanek

And it's interesting that during the first pass on a dirty image it works well even on the slave too. -- Pavel 2013/9/27 Pavel Krivanek <pavel.krivanek(a)gmail.com>: > It fails during Release test. > > Smalltalk stack dump: > 0xbfee7364 M CompiledMethod class(Behavior)>allInstances 0x776a94b4: > a(n) CompiledMethod class > 0xbfee7380 M SmalltalkImage>fixObsoleteReferences 0x77916420: a(n) > SmalltalkImage > 0xbfee5234 M ReleaseTest>testObsoleteClasses 0x78e70cac: a(n) ReleaseTest > > I tried it on my local Linux machine but it worked well :-/ > > -- Pavel > > 2013/9/27 Marcus Denker <marcus.denker(a)inria.fr>: >> >> On Sep 26, 2013, at 8:59 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: >> >>> The update build failed on Linux slave >>> >> a second run leads to the same problem. This is not goodâ¦. >> >>> -- Pavel >>> >>> 2013/9/26 Marcus Denker <marcus.denker(a)inria.fr>: >>>> 30425 >>>> ----- >>>> >>>> 11687 Merge Opal with Opal Repository >>>> https://pharo.fogbugz.com/f/cases/11687 >>>> >>>> 11688 Pharo Kernel shrinking update for 30424 >>>> https://pharo.fogbugz.com/f/cases/11688 >>>> >>>> Diff information: >>>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/System-Support-MarcusDenker.9… >>>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Tests-MarcusDenk… >>>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Core-MarcusDenke… >>>> >>> >>

Sept. 27, 2013

Re: [Pharo-dev] [update 3.0] #30425

by Pavel Krivanek

It fails during Release test. Smalltalk stack dump: 0xbfee7364 M CompiledMethod class(Behavior)>allInstances 0x776a94b4: a(n) CompiledMethod class 0xbfee7380 M SmalltalkImage>fixObsoleteReferences 0x77916420: a(n) SmalltalkImage 0xbfee5234 M ReleaseTest>testObsoleteClasses 0x78e70cac: a(n) ReleaseTest I tried it on my local Linux machine but it worked well :-/ -- Pavel 2013/9/27 Marcus Denker <marcus.denker(a)inria.fr>: > > On Sep 26, 2013, at 8:59 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > >> The update build failed on Linux slave >> > a second run leads to the same problem. This is not goodâ¦. > >> -- Pavel >> >> 2013/9/26 Marcus Denker <marcus.denker(a)inria.fr>: >>> 30425 >>> ----- >>> >>> 11687 Merge Opal with Opal Repository >>> https://pharo.fogbugz.com/f/cases/11687 >>> >>> 11688 Pharo Kernel shrinking update for 30424 >>> https://pharo.fogbugz.com/f/cases/11688 >>> >>> Diff information: >>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/System-Support-MarcusDenker.9… >>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Tests-MarcusDenk… >>> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Core-MarcusDenke… >>> >> >

Sept. 27, 2013

Re: [Pharo-dev] [update 3.0] #30425

by Marcus Denker

On Sep 26, 2013, at 8:59 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > The update build failed on Linux slave > a second run leads to the same problem. This is not goodâ¦. > -- Pavel > > 2013/9/26 Marcus Denker <marcus.denker(a)inria.fr>: >> 30425 >> ----- >> >> 11687 Merge Opal with Opal Repository >> https://pharo.fogbugz.com/f/cases/11687 >> >> 11688 Pharo Kernel shrinking update for 30424 >> https://pharo.fogbugz.com/f/cases/11688 >> >> Diff information: >> http://smalltalkhub.com/mc/Pharo/Pharo30/main/System-Support-MarcusDenker.9… >> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Tests-MarcusDenk… >> http://smalltalkhub.com/mc/Pharo/Pharo30/main/OpalCompiler-Core-MarcusDenke… >> >

Sept. 27, 2013

[regression reporter]regression occurred

by no-reply＠ci.inria.fr

https://ci.inria.fr/pharo/job/Pharo-3.0-Update-Step-2.1-Validation/label=wi… 2 regressions found. Zinc.Tests.ZnServerTests.testEntityTooLarge Zinc.Zodiac.ZnHTTPSTests.testGetPharoVersion

Sept. 27, 2013

Re: [Pharo-dev] SmalltalkHub use of LESS stylesheets

by Sebastian Sastre

I'm curions on how this develops. Looking forward for hear about every tactic to make amber apps loading snappishly :D On Sep 26, 2013, at 11:27 PM, Esteban A. Maringolo <emaringolo(a)gmail.com> wrote: > Hi all, > > The SmalltalkHub website seems to be using LESS stylesheets, which are > downloaded in a sequential order, and it takes ~20sec to download the > styles. > > Wouldn't it be better to compile the LESS stylesheets to regular CSS > files (combined and minified in a single file would be great). LESS > stylesheets are not meant to be used in production environments. > > I thinks that only with that the speedup will be significant. > > Regards, > > Esteban A. Maringolo >

Sept. 27, 2013