April 2021 - Pharo-dev - lists.pharo.org

Re: Improving Variable Shadow model

by Marcus Denker

Just a status update: we merged all the code. - we now have for both the class and method view a Critique rule that warns you if you use shadowing vars. - Pavel fixed all the cases of shadowing class variables. Thanks! - release tests are active, #testNoShadowedVariablesInMethods fo methods and at the class level: testClassesShadow | classes | classes := Smalltalk globals allBehaviors select: [ :class | class definedVariables anySatisfy: [ :var | var isShadowing ] ]. self assert: classes isEmpty description: classes asArray asString status cleanups: - DONE: Shadowed vars (as we saw) => TODO: combine the snippets of writing into one blog post for the devblog - DONE: unused Class Variables - ongoing: unused ivars. Just some left in Spec+Newtools (and one in Kernel tests) - ongoing: all methods should be categorized. - todo: we should have no method that is the same as a method in a superclass > On 2 Apr 2021, at 15:33, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > > > # Code Critique > > What we need next is a Code Critique rule. This for once warns developers early, but in addition it will > make it much easier to fix the 6 problem cases that the release test exposed. > > Checking for current senders of #definedVariables leads us to ReVariableAssignedLiteralRule, we can use this as a template. > The main method doing the check looks like this: > > ```Smalltalk > check: aClass forCritiquesDo: aCriticBlock > > aClass definedVariables do: [ :variable | > variable isShadowing ifTrue: [ > aCriticBlock cull: (self critiqueFor: aClass about: variable name) ] ] > ``` > > The full PR is here: https://github.com/pharo-project/pharo/pull/8940 <https://github.com/pharo-project/pharo/pull/8940> > >

April 14, 2021

Progress Report -> Refactoring Project - ( April 5 - April 9 )

by Evelyn Cusi Lopez

Hi everyone, Last week I did these tasks: - *Add a default refactoring scope. *Now you can define a default environment through Refactorings Settings, this is useful when you are doing refactoring changes in Global environment. You can see the new behavior in this video: https://www.youtube.com/watch?v=NkxGP0P9klA . Check PR <https://github.com/pharo-project/pharo/pull/8999> to know the changes. - *Improve "Set up score command" *Check PR <https://github.com/pharo-project/pharo/pull/9039> to know the changes. -* Fix move method with empty protocol* (Move to package of defining class enh #8929 <https://github.com/pharo-project/pharo/issues/8929>) Check PR <https://github.com/pharo-project/pharo/pull/9040> to know the changes. - *Add preview rename class refactoring* Check PR <https://github.com/pharo-project/pharo/pull/9043> to know the changes. - I was fixing the issue: https://github.com/pharo-project/pharo/issues/6218 but I found two issues more, these are: RBNamespace doesn't support alias #9041 <https://github.com/pharo-project/pharo/issues/9041> , Inline method refactoring doesn't handle well, handling calls to super or overrides #9042 <https://github.com/pharo-project/pharo/issues/9042> Still in progress. - I was improving the creation of cascades to fix: Inline temp refactoring can create code with different behavior. #6056 <https://github.com/pharo-project/pharo/issues/6056>Still in progress. Tasks for this week: - Fix issue: RBNamespace doesn't support alias #9041 <https://github.com/pharo-project/pharo/issues/9041> - Fix issue: Inline method refactoring doesn't handle well, handling calls to super or overrides #9042 <https://github.com/pharo-project/pharo/issues/9042> - Fix issue: Inline senders refactoring does not correctly manage errors #6218 <https://github.com/pharo-project/pharo/issues/6218> - Finish to fix: Create cascade issue <https://github.com/pharo-project/pharo/issues/6056>. (Fix when the object is referenced after cascade block, fix the assignation when we have returns in called methods) - Fix issue: Introduce class Deprecation as Refactoring #9033 <https://github.com/pharo-project/pharo/issues/9033> <https://github.com/pharo-project/pharo/issues/8871> - Fix issue: move to component refactoring issue. <https://github.com/pharo-project/pharo/issues/8499> <https://github.com/pharo-project/pharo/issues/6218> Regards, Evelyn C.

April 12, 2021

Re: Pharo 8.0 crash on resizing with Moom

by Sven Van Caekenberghe

Hi Bernard, I run a dual monitor macOS based setup as well. In general there is indeed a hard to track down issue with going full screen. What does work for me is going 'almost' full screen, that is, leave the menu bar visible. You can do that with Option-click on the green expand button. I think this is called Zoom. That way everything remains rock solid for me (as in images are running for days on end). I like this better than real full screen because the menu bar has useful functionality (especially the right hand side icons), even though Pharo does not need or use it. HTH, Sven > On 10 Apr 2021, at 11:46, Bernhard Pieber <bernhard(a)pieber.com> wrote: > > Hi everyone, > > Is this the right place to report a Pharo 8.0 crash? > > I am on a MacBookPro16,4 on macOS Big Sur Version 11.2.3 (20D91) with an external 30" Apple Cinema Display (2500 x 1600). > > Here is what I did: > > I downloaded PharoLauncher. > > I created a new image using the official distribution Pharo 8.0 - 64bit (stable). > > I launched the new image. > > I clicked on light theme. > > I resized to full screen using Moom. => It crashed. > > I could reproduce it three out of five times I tried those steps. > > Let me add that Moom works very reliably with all other applications. Resizing like that is in my muscular memory. > > Cheers, > Bernhard > I dow > <crash.dmp>

April 11, 2021

Pharo 8.0 crash on resizing with Moom

by Bernhard Pieber

Hi everyone, Is this the right place to report a Pharo 8.0 crash? I am on a MacBookPro16,4 on macOS Big Sur Version 11.2.3 (20D91) with an external 30" Apple Cinema Display (2500 x 1600). Here is what I did: I downloaded PharoLauncher. I created a new image using the official distribution Pharo 8.0 - 64bit (stable). I launched the new image. I clicked on light theme. I resized to full screen using Moom. => It crashed. I could reproduce it three out of five times I tried those steps. Let me add that Moom works very reliably with all other applications. Resizing like that is in my muscular memory. Cheers, Bernhard I dow

April 10, 2021

This week (14/2021) on the Pharo Issue Tracker

by Marcus Denker

We merged 36 Pull Requests and closed 29 issues. https://github.com/pharo-project/pharo/pulse Merge newtools 0.6.2 #8951 ========================== - https://github.com/pharo-spec/NewTools/releases/tag/v0.6.2 - enhance control of spotter queries in a presenter (it should present "old" results to appear in new queries) - fixes pharo-spec/NewTools#209 - fixes pharo-spec/NewTools#196 - using a worker pool instead a regular "schedule" - fix a problem of entries being duplicated on history Spec 0.8.8 =========== - https://github.com/pharo-spec/Spec/releases/tag/v0.8.8 - "labelled" presenters moved to deprecated and not loaded by default. - tweak how overlay stack (still can be wrong) - fix a problem when assigning box layout sizes : it was not stacking widget sizes. Refactoring Tools ================ - RB - Method's Refactoring using selected scope #8950 https://github.com/pharo-project/pharo/pull/8950 - RB - Correct selection scope #8960 https://github.com/pharo-project/pharo/pull/8960 - RB - Clean use of "Smalltalk globals" #8959 https://github.com/pharo-project/pharo/pull/8959 Shadowing Variables =================== https://blog.marcusdenker.de/modeling-variable-shadowing-part-ii https://blog.marcusdenker.de/modeling-variable-shadowing-part-iii - Shadowing-Better-Implementation #8931 https://github.com/pharo-project/pharo/pull/8931 - Simplify-OCShadowVariableWarning #8943 https://github.com/pharo-project/pharo/pull/8943 - ReClassVariablesShadowingRule #8940 https://github.com/pharo-project/pharo/pull/8940 - ReClassVariablesShadowingRule-Improve #8965 https://github.com/pharo-project/pharo/pull/8965 - further improve testUndeclared #8964 https://github.com/pharo-project/pharo/pull/8964 Cleanup Shadowing Vars ======================= With the model for shadowing vars merged, we can cleanup. - LocalTimeZone class variable is shadowing the class of the same name. #8979 https://github.com/pharo-project/pharo/pull/8979 - ReleaseTest: testNoShadowedVariablesInMethods #8927 https://github.com/pharo-project/pharo/pull/8927 - reset the ASTCache in #testNoShadowedVariablesInMethods #8986 https://github.com/pharo-project/pharo/pull/8986 - 8935-add-release-test-for-shadowed-ivar--class-vars #8939 https://github.com/pharo-project/pharo/pull/8939 - ClassVariable in RBBrowserEnvironmentTest is shadowing a class #8982 https://github.com/pharo-project/pharo/pull/8982 - ClassVariable in OCOpalExamples is shadowing a class #8980 https://github.com/pharo-project/pharo/pull/8980 - Author in PoolDefiner and PoolDefiner2 is shadowing class #8981 https://github.com/pharo-project/pharo/pull/8981 - Link in ReflectivityDemos is shadowing a class #8983 https://github.com/pharo-project/pharo/pull/8983 - 8988-Failing-tests-ClassTest-testUsesPoolVarNamed-and-testHasPoolVarNamed- #8989 https://github.com/pharo-project/pharo/pull/8989 Fixes ===== - allow system to react to a theme change #8937 https://github.com/pharo-project/pharo/pull/8937 - 8861-Code-Coverage-in-Dr-Tests-is-misreporting-methods-implemented-as-subclassResponsibility #8954 https://github.com/pharo-project/pharo/pull/8954 - 7655-Epicea-fileout-should-produce-Pharo-code-and-not-ombu-definition #8969 https://github.com/pharo-project/pharo/pull/8969 - Don't make SDL_Texture autoRelease from renderer #8962 https://github.com/pharo-project/pharo/pull/8962 - ZdcSecureSocketStreamTest>>#testPlain and ZnHTTPSTests>>#testRequestRâ¦ #8977 https://github.com/pharo-project/pharo/pull/8977 - 8972-Class-instance-variable-menu-error #8973 https://github.com/pharo-project/pharo/pull/8973 - Float numbers are only immediates if the VM/Image has SmallFloats. #8975 https://github.com/pharo-project/pharo/pull/8975 - Use a Mutex instead of a Semaphore to prevent trouble in concurrent sâ¦ #8990 https://github.com/pharo-project/pharo/pull/8990 - CairoLibrary in windows: don't look for cairo.dll #8994 https://github.com/pharo-project/pharo/pull/8994 - fixes: #7599 finally class side fluid does not show anymore the brokeâ¦ #8987 https://github.com/pharo-project/pharo/pull/8987 Cleanups ======== - AST-Tag-ErrorNodes #8936 https://github.com/pharo-project/pharo/pull/8936 - 8932-CI-failure-ArrayTestClassTestCasetestMethodsOfTheClassShouldNotBeRepeatedInItsSuperclasses https://github.com/pharo-project/pharo/pull/8938 - #8942 another try: Cleanup some class name in Drtests #8948 https://github.com/pharo-project/pharo/pull/8948 - 8955-failing-test-on-CI-ReleaseTesttestUndeclared #8958 https://github.com/pharo-project/pharo/pull/8958 - CI - Fix broken test ReleaseTest >> #testUndeclared #8971 https://github.com/pharo-project/pharo/pull/8971 - 5388-Improving-RBNode-API: add #hasArguments #8949 https://github.com/pharo-project/pharo/pull/8949 - RBSequenceNode-children-no-accessor #8963 https://github.com/pharo-project/pharo/pull/8963 - 7827-oldDefinition-should-use-the-associated-class-definition-printer #8970 https://github.com/pharo-project/pharo/pull/8970

April 9, 2021

Re: Broken window alert

by tesonep＠gmail.com

Thanks, it looks cool On Wed, Apr 7, 2021 at 12:12 PM Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > And here is the PR, I hope it is OK: > > https://github.com/pharo-project/pharo/pull/8977 > > > On 7 Apr 2021, at 11:53, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > > > Hi Pablo, Guille, > > > > Thanks for the links. > > > > You can consider the Zn/Zdc failures fixed. Something changed recently > with the Google query service so that it now includes a redirect. The full > ZnClient follows those without any problems, but these tests are much more > primitive and can't handle a redirect. I now use DuckDuckGo for the test at > this level. > > > > Here is the first change > > > > > https://github.com/svenvc/zinc/commit/75710f9832e6ddad8a477d061978149b1d08c… > > > > and the second > > > > > https://github.com/svenvc/zodiac/commit/cd4c4ee7e7cc8eb3624b4c7b4f8af861481… > > > > my own tests for Pharo 7, 8 and 9 are green > > > > https://github.com/svenvc/zinc/actions/runs/725549812 > > > > I will now try to make a PR on a current Pharo 9 image. > > > > Sven > > > >> On 6 Apr 2021, at 22:58, Guillermo Polito <guillermopolito(a)gmail.com> > wrote: > >> > >> > >> > >>> El 6 abr 2021, a las 22:55, Sven Van Caekenberghe <sven(a)stfx.eu> > escribiÃ³: > >>> > >>> I agree, could you post a link to or a list of the broken tests ? > >> > >> Sure :) > >> > >> The CI: > >> > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > >> > >> Last Three Jobs: > >> > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > >> > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > >> > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > >> > >> > >>> > >>>> On 6 Apr 2021, at 22:54, Guillermo Polito <guillermopolito(a)gmail.com> > wrote: > >>>> > >>>> Hi Guys, > >>>> > >>>> With Pablo we were discussing that it would be nice to stop with the > integrations until we fix the broken tests. > >>>> We have crossed the 20 broken tests for more than 20 builds. > >>>> We should put some energy in fixing those, but we are afraid that if > we continue this way, eventually it will become unmanageable. > >>>> > >>>> Cheers, > >>>> Guille > >> > > > -- Pablo Tesone. tesonep(a)gmail.com

April 7, 2021

Re: Broken window alert

by Sven Van Caekenberghe

And here is the PR, I hope it is OK: https://github.com/pharo-project/pharo/pull/8977 > On 7 Apr 2021, at 11:53, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > Hi Pablo, Guille, > > Thanks for the links. > > You can consider the Zn/Zdc failures fixed. Something changed recently with the Google query service so that it now includes a redirect. The full ZnClient follows those without any problems, but these tests are much more primitive and can't handle a redirect. I now use DuckDuckGo for the test at this level. > > Here is the first change > > https://github.com/svenvc/zinc/commit/75710f9832e6ddad8a477d061978149b1d08c… > > and the second > > https://github.com/svenvc/zodiac/commit/cd4c4ee7e7cc8eb3624b4c7b4f8af861481… > > my own tests for Pharo 7, 8 and 9 are green > > https://github.com/svenvc/zinc/actions/runs/725549812 > > I will now try to make a PR on a current Pharo 9 image. > > Sven > >> On 6 Apr 2021, at 22:58, Guillermo Polito <guillermopolito(a)gmail.com> wrote: >> >> >> >>> El 6 abr 2021, a las 22:55, Sven Van Caekenberghe <sven(a)stfx.eu> escribiÃ³: >>> >>> I agree, could you post a link to or a list of the broken tests ? >> >> Sure :) >> >> The CI: >> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… >> >> Last Three Jobs: >> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… >> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… >> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… >> >> >>> >>>> On 6 Apr 2021, at 22:54, Guillermo Polito <guillermopolito(a)gmail.com> wrote: >>>> >>>> Hi Guys, >>>> >>>> With Pablo we were discussing that it would be nice to stop with the integrations until we fix the broken tests. >>>> We have crossed the 20 broken tests for more than 20 builds. >>>> We should put some energy in fixing those, but we are afraid that if we continue this way, eventually it will become unmanageable. >>>> >>>> Cheers, >>>> Guille >> >

April 7, 2021

Re: Broken window alert

by Sven Van Caekenberghe

Hi Pablo, Guille, Thanks for the links. You can consider the Zn/Zdc failures fixed. Something changed recently with the Google query service so that it now includes a redirect. The full ZnClient follows those without any problems, but these tests are much more primitive and can't handle a redirect. I now use DuckDuckGo for the test at this level. Here is the first change https://github.com/svenvc/zinc/commit/75710f9832e6ddad8a477d061978149b1d08c… and the second https://github.com/svenvc/zodiac/commit/cd4c4ee7e7cc8eb3624b4c7b4f8af861481… my own tests for Pharo 7, 8 and 9 are green https://github.com/svenvc/zinc/actions/runs/725549812 I will now try to make a PR on a current Pharo 9 image. Sven > On 6 Apr 2021, at 22:58, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > > > >> El 6 abr 2021, a las 22:55, Sven Van Caekenberghe <sven(a)stfx.eu> escribiÃ³: >> >> I agree, could you post a link to or a list of the broken tests ? > > Sure :) > > The CI: > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > > Last Three Jobs: > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… > > >> >>> On 6 Apr 2021, at 22:54, Guillermo Polito <guillermopolito(a)gmail.com> wrote: >>> >>> Hi Guys, >>> >>> With Pablo we were discussing that it would be nice to stop with the integrations until we fix the broken tests. >>> We have crossed the 20 broken tests for more than 20 builds. >>> We should put some energy in fixing those, but we are afraid that if we continue this way, eventually it will become unmanageable. >>> >>> Cheers, >>> Guille >

April 7, 2021

Re: Broken window alert

by Guillermo Polito

> El 6 abr 2021, a las 22:55, Sven Van Caekenberghe <sven(a)stfx.eu> escribiÃ³: > > I agree, could you post a link to or a list of the broken tests ? Sure :) The CI: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… <https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…> Last Three Jobs: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… <https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… <https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…> https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… <https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…> > >> On 6 Apr 2021, at 22:54, Guillermo Polito <guillermopolito(a)gmail.com> wrote: >> >> Hi Guys, >> >> With Pablo we were discussing that it would be nice to stop with the integrations until we fix the broken tests. >> We have crossed the 20 broken tests for more than 20 builds. >> We should put some energy in fixing those, but we are afraid that if we continue this way, eventually it will become unmanageable. >> >> Cheers, >> Guille

April 6, 2021

Re: Broken window alert

by tesonep＠gmail.com

Hi Sven, Here there is: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2… Thanks, Pablo On Tue, Apr 6, 2021 at 10:56 PM Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > I agree, could you post a link to or a list of the broken tests ? > > > On 6 Apr 2021, at 22:54, Guillermo Polito <guillermopolito(a)gmail.com> > wrote: > > > > Hi Guys, > > > > With Pablo we were discussing that it would be nice to stop with the > integrations until we fix the broken tests. > > We have crossed the 20 broken tests for more than 20 builds. > > We should put some energy in fixing those, but we are afraid that if we > continue this way, eventually it will become unmanageable. > > > > Cheers, > > Guille > -- Pablo Tesone. tesonep(a)gmail.com

April 6, 2021