- Pharo-dev - lists.pharo.org

Missing tests for FFI in CI

by Stephan Eggermont

Yesterday, we tracked down a serious issue in 64 bit FFI. After giving up with our own code, we made a fresh start with the latest base image and OSWindows. We were finally able to figure out what went wrong with two simple tests that were failing in 64 bit that worked correctly in 32 bit. FFIConstantHandleType externalType is the wrong size After confirming that and making it work, we were told that a number of FFI fixes were done by Nicolas and Eliot in Squeak that did not get integrated in Pharo. Applying the changes from the latest Squeak FFI resolved the problem, and a few that we were unaware of. Stephan

March 21, 2019

Re: [Pharo-dev] Why everybody could have an impact?

by Tim Mackinnon

Yes, I got to do it (so he could help in a small way) - https://github.com/pharo-project/pharo/issues/2939 And I annotated it. Tim Sent from my iPhone > On 20 Mar 2019, at 20:08, ducasse <stepharo(a)netcourrier.com> wrote: > > Hi tim > > Can you report it precisely because right now it is not actionable. > And yes doing something is a challenge. Doing nothing is much more confortable because nothing breaks, and you die relaxed. > > Stef > >> On 20 Mar 2019, at 20:17, Tim Mackinnon <tim(a)testit.works> wrote: >> >> Ironically - one of my Exercism testers hit an issue with extract method about when you typed this... itâs seems there has a difference between âSuggestions | Extract methodâ and âSource Code | Extract methodâ sigh... the former gives a walk back ... and itâs due to a strange misspelled variable âpreviousSelectionHighligthâ not being initialised. >> >> So yeah - this all needs testing and fixing. >> >> Tim >> >> Sent from my iPhone >> >>> On 20 Mar 2019, at 17:16, ducasse <stepharo(a)netcourrier.com> wrote: >>> >>> Hi >>> >>> to relax I was going over RB code because we will start to improve the refactoring. >>> And I started to add tests (yes the dummy little things that everybody can write but >>> that most people prefer to think they do and talk about). >>> And writing such super super stupid tests >>> >>> testCheckInvalidMethodName >>> "Usually used to validate input." >>> >>> self deny: (RBCondition checkMethodName: 'fofo fo'). >>> self deny: (RBCondition checkMethodName: '123fofo'). >>> "self deny: (RBCondition checkMethodName: 'foo::')." >>> "self deny: (RBCondition checkMethodName: 'agr:goo:aa').â >>> >>> >>> checkMethodName: aString >>> "Return whether the argument aName is can represent a selector" >>> >>> ^ aString isString and: [ RBScanner isSelector: aString ] >>> >>> >>> I found that RBScanner reports that >>> >>> #foo:: or 'agr:goo:aaâ is a valid selector :( >>> >>> So if you **really** want to help pharo this is not that difficult. >>> >>> Now this is a matter of will. >>> >>> Stef >>> >> >> > > >

March 20, 2019

Re: [Pharo-dev] Why everybody could have an impact?

by ducasse

Hi tim Can you report it precisely because right now it is not actionable. And yes doing something is a challenge. Doing nothing is much more confortable because nothing breaks, and you die relaxed. Stef > On 20 Mar 2019, at 20:17, Tim Mackinnon <tim(a)testit.works> wrote: > > Ironically - one of my Exercism testers hit an issue with extract method about when you typed this... itâs seems there has a difference between âSuggestions | Extract methodâ and âSource Code | Extract methodâ sigh... the former gives a walk back ... and itâs due to a strange misspelled variable âpreviousSelectionHighligthâ not being initialised. > > So yeah - this all needs testing and fixing. > > Tim > > Sent from my iPhone > >> On 20 Mar 2019, at 17:16, ducasse <stepharo(a)netcourrier.com> wrote: >> >> Hi >> >> to relax I was going over RB code because we will start to improve the refactoring. >> And I started to add tests (yes the dummy little things that everybody can write but >> that most people prefer to think they do and talk about). >> And writing such super super stupid tests >> >> testCheckInvalidMethodName >> "Usually used to validate input." >> >> self deny: (RBCondition checkMethodName: 'fofo fo'). >> self deny: (RBCondition checkMethodName: '123fofo'). >> "self deny: (RBCondition checkMethodName: 'foo::')." >> "self deny: (RBCondition checkMethodName: 'agr:goo:aa').â >> >> >> checkMethodName: aString >> "Return whether the argument aName is can represent a selector" >> >> ^ aString isString and: [ RBScanner isSelector: aString ] >> >> >> I found that RBScanner reports that >> >> #foo:: or 'agr:goo:aaâ is a valid selector :( >> >> So if you **really** want to help pharo this is not that difficult. >> >> Now this is a matter of will. >> >> Stef >> > >

March 20, 2019

Re: [Pharo-dev] About a better filedialog

by Esteban Lorenzano

Hi, While a good FileDialog may have a previewâ¦ I think the best thing is that you can take the component and reuse it, adding the behaviour you want. At least, that would be the purpose of Spec, isnât? Soâ¦ you could not wait until someone add a preview and you could add the preview you like it :) Esteban Ps: Of course, I know that means timeâ¦ and I know a new component has to provide the same functionalities as the older one at least. I was just pointing it :) > On 20 Mar 2019, at 20:11, ducasse <stepharo(a)netcourrier.com> wrote: > > I checked and they are not related. > And we can keep for now the DileDialogWindow. > > Stef > >> On 20 Mar 2019, at 20:09, ducasse <stepharo(a)netcourrier.com> wrote: >> >> >> >>> On 20 Mar 2019, at 19:29, Hilaire <hilaire(a)drgeo.eu> wrote: >>> >>> Ah, may be we are not talking about the same tool. I am talking about >>> the FileDialogWindow, I think part of the Gary's Polymorph. And you >>> about FileList according to your screenshot. >>> >>> Is Peter's FileDialog intend is to replace FileList? Or FileDialogWindow? >> >> I was thinking the FileList. >> Now I will look at the FileDialogWindow :) >> >> Stef >>> >>> Hilaire >>> >>> >>> Le 20/03/2019 Ã 19:13, ducasse a Ã©crit : >>>> Ok I see and how the preview is done? >>>> You register some objects to the fileDialog? >>>> Because I get this one. >>>> >>>> So may be we are talking about a different one. >>>> >>> -- >>> Dr. Geo >>> http://drgeo.eu >>> >>> >>> >> >> >> > > >

March 20, 2019

Re: [Pharo-dev] Why everybody could have an impact?

by Tim Mackinnon

Ironically - one of my Exercism testers hit an issue with extract method about when you typed this... itâs seems there has a difference between âSuggestions | Extract methodâ and âSource Code | Extract methodâ sigh... the former gives a walk back ... and itâs due to a strange misspelled variable âpreviousSelectionHighligthâ not being initialised. So yeah - this all needs testing and fixing. Tim Sent from my iPhone > On 20 Mar 2019, at 17:16, ducasse <stepharo(a)netcourrier.com> wrote: > > Hi > > to relax I was going over RB code because we will start to improve the refactoring. > And I started to add tests (yes the dummy little things that everybody can write but > that most people prefer to think they do and talk about). > And writing such super super stupid tests > > testCheckInvalidMethodName > "Usually used to validate input." > > self deny: (RBCondition checkMethodName: 'fofo fo'). > self deny: (RBCondition checkMethodName: '123fofo'). > "self deny: (RBCondition checkMethodName: 'foo::')." > "self deny: (RBCondition checkMethodName: 'agr:goo:aa').â > > > checkMethodName: aString > "Return whether the argument aName is can represent a selector" > > ^ aString isString and: [ RBScanner isSelector: aString ] > > > I found that RBScanner reports that > > #foo:: or 'agr:goo:aaâ is a valid selector :( > > So if you **really** want to help pharo this is not that difficult. > > Now this is a matter of will. > > Stef >

March 20, 2019

Re: [Pharo-dev] About a better filedialog

by ducasse

I checked and they are not related. And we can keep for now the DileDialogWindow. Stef > On 20 Mar 2019, at 20:09, ducasse <stepharo(a)netcourrier.com> wrote: > > > >> On 20 Mar 2019, at 19:29, Hilaire <hilaire(a)drgeo.eu> wrote: >> >> Ah, may be we are not talking about the same tool. I am talking about >> the FileDialogWindow, I think part of the Gary's Polymorph. And you >> about FileList according to your screenshot. >> >> Is Peter's FileDialog intend is to replace FileList? Or FileDialogWindow? > > I was thinking the FileList. > Now I will look at the FileDialogWindow :) > > Stef >> >> Hilaire >> >> >> Le 20/03/2019 Ã 19:13, ducasse a Ã©crit : >>> Ok I see and how the preview is done? >>> You register some objects to the fileDialog? >>> Because I get this one. >>> >>> So may be we are talking about a different one. >>> >> -- >> Dr. Geo >> http://drgeo.eu >> >> >> > > >

March 20, 2019

Re: [Pharo-dev] About a better filedialog

by ducasse

> On 20 Mar 2019, at 19:29, Hilaire <hilaire(a)drgeo.eu> wrote: > > Ah, may be we are not talking about the same tool. I am talking about > the FileDialogWindow, I think part of the Gary's Polymorph. And you > about FileList according to your screenshot. > > Is Peter's FileDialog intend is to replace FileList? Or FileDialogWindow? I was thinking the FileList. Now I will look at the FileDialogWindow :) Stef > > Hilaire > > > Le 20/03/2019 Ã 19:13, ducasse a Ã©crit : >> Ok I see and how the preview is done? >> You register some objects to the fileDialog? >> Because I get this one. >> >> So may be we are talking about a different one. >> > -- > Dr. Geo > http://drgeo.eu > > >

March 20, 2019

Re: [Pharo-dev] About a better filedialog

by Hilaire

Ah, may be we are not talking about the same tool. I am talking about the FileDialogWindow, I think part of the Gary's Polymorph. And you about FileList according to your screenshot. Is Peter's FileDialog intend is to replace FileList? Or FileDialogWindow? Hilaire Le 20/03/2019 Ã 19:13, ducasse a Ã©critÂ : > Ok I see and how the preview is done? > You register some objects to the fileDialog? > Because I get this one.Â > > So may be we are talking about a different one. > -- Dr. Geo http://drgeo.eu

March 20, 2019

Re: [Pharo-dev] About a better filedialog

by ducasse

Ok I see and how the preview is done? You register some objects to the fileDialog? Because I get this one. So may be we are talking about a different one. > On 20 Mar 2019, at 18:16, Hilaire <hilaire(a)drgeo.eu> wrote: > > See this filedialog screenshot. > > Hilaire > > -- > Dr. Geo > http://drgeo.eu > > <Choisir un fichier Dr. Geo.png>

March 20, 2019

Why everybody could have an impact?

by ducasse

Hi to relax I was going over RB code because we will start to improve the refactoring. And I started to add tests (yes the dummy little things that everybody can write but that most people prefer to think they do and talk about). And writing such super super stupid tests testCheckInvalidMethodName "Usually used to validate input." self deny: (RBCondition checkMethodName: 'fofo fo'). self deny: (RBCondition checkMethodName: '123fofo'). "self deny: (RBCondition checkMethodName: 'foo::')." "self deny: (RBCondition checkMethodName: 'agr:goo:aa').â checkMethodName: aString "Return whether the argument aName is can represent a selector" ^ aString isString and: [ RBScanner isSelector: aString ] I found that RBScanner reports that #foo:: or 'agr:goo:aaâ is a valid selector :( So if you **really** want to help pharo this is not that difficult. Now this is a matter of will. Stef

March 20, 2019