October 2015 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] #50409 regression

by stepharo

> >> >> 2015-10-28 16:19 GMT+01:00 stepharo <stepharo(a)free.fr >> <mailto:stepharo@free.fr>>: >> >> I see this is annoying that comments are not polymorphic. >> >> >> They could be :) >> - an empty RBMethodNode with a comment attribute, >> or >> - a comment: message send with the comment as a parameter. > > Would be nice to have a tree node > > ClassComment > > > You have RGCommentDefinition, no? And RBComment So we should probably use them. > > > The static representation of Pharo program can really be improved. > > I still want a full package. > > > RGPackage? Or a unified AST-like structure? Even an AST is not that > cool as a program representation. > I'm thorn because on one hand I would like to be able to express a visitor for the complete compilation unit and in particular package and extensions but may be we do not need to have this nodes and just have an object. I would be favor to have a nice and full AST representation of the compilation unit. Stef

Oct. 29, 2015

Re: [Pharo-dev] #50409 regression

by Thierry Goubier

2015-10-29 14:58 GMT+01:00 stepharo <stepharo(a)free.fr>: > > > 2015-10-28 16:19 GMT+01:00 stepharo <stepharo(a)free.fr>: > >> I see this is annoying that comments are not polymorphic. >> > > They could be :) > - an empty RBMethodNode with a comment attribute, > or > - a comment: message send with the comment as a parameter. > > > Would be nice to have a tree node > > ClassComment > You have RGCommentDefinition, no? > > The static representation of Pharo program can really be improved. > I still want a full package. > RGPackage? Or a unified AST-like structure? Even an AST is not that cool as a program representation. Thierry > > Stef > > > > Thierry > > >> >> Stef >> >> Le 28/10/15 11:05, Nicolai Hess a Ã©crit : >> >> >> >> 2015-10-28 15:33 GMT+01:00 stepharo < <stepharo(a)free.fr>stepharo(a)free.fr> >> : >> >>> Hi >>> >>> I typed 'Take' and source code with it. >>> - the progress bar does not update >>> >>> - more important when I click on the message browser I get a DNU ast >>> >>> autoSelect: aSelector >>> textModel readSelectionBlock: [ :text | >>> self selectedMessage >>> ifNotNil: [ :msg | >>> (self sourceIntervalOf: aSelector in: msg ast) >>> ^^^^^^^^^^^^^ >>> ifEmpty:[ self searchedString: aSelector >>> asString in: text ] ]] >>> >>> and RGCommentDefinition does not understand ast. >>> So what do we do? >>> >> >> It is a know bug ( at least, I know of it :) ) >> I changed the way the MessageBrowser selects the search term, >> because it always highlights the search term by "text" and not by "code" >> ,for example, in a "at:put:" senders MessageBrowser >> the first the highlighting started at the first >> at: >> although this is part of a at:ifAbsent: - message. >> >> at: #OneShotArmed ifAbsent: [ true ]) >> ifTrue: [ >> Smalltalk globals at: #OneShotArmed put: >> >> 16673 >> <https://pharo.fogbugz.com/f/cases/16673/Senders-of-highlights-wrong-code-se…> >> Senders of... highlights wrong code segment >> >> But I didn't take the "class-comment"-search results into account. >> >> >> >> >> >> >>> >>> Is it a well know bug? >>> >>> Stef >>> >>> >> >> > >

Oct. 29, 2015

Re: [Pharo-dev] #50409 regression

by stepharo

> > 2015-10-28 16:19 GMT+01:00 stepharo <stepharo(a)free.fr > <mailto:stepharo@free.fr>>: > > I see this is annoying that comments are not polymorphic. > > > They could be :) > - an empty RBMethodNode with a comment attribute, > or > - a comment: message send with the comment as a parameter. Would be nice to have a tree node ClassComment The static representation of Pharo program can really be improved. I still want a full package. Stef > > Thierry > > > Stef > > Le 28/10/15 11:05, Nicolai Hess a Ã©crit : >> >> >> 2015-10-28 15:33 GMT+01:00 stepharo <stepharo(a)free.fr >> <mailto:stepharo@free.fr>>: >> >> Hi >> >> I typed 'Take' and source code with it. >> - the progress bar does not update >> >> - more important when I click on the message browser I >> get a DNU ast >> >> autoSelect: aSelector >> textModel readSelectionBlock: [ :text | >> self selectedMessage >> ifNotNil: [ :msg | >> (self sourceIntervalOf: aSelector in: msg ast) >> ^^^^^^^^^^^^^ >> ifEmpty:[ self searchedString: >> aSelector asString in: text ] ]] >> >> and RGCommentDefinition does not understand ast. >> So what do we do? >> >> >> It is a know bug ( at least, I know of it :) ) >> I changed the way the MessageBrowser selects the search term, >> because it always highlights the search term by "text" and not by >> "code" ,for example, in a "at:put:" senders MessageBrowser >> the first the highlighting started at the first >> at: >> although this is part of a at:ifAbsent: - message. >> >> at: #OneShotArmed ifAbsent: [ true ]) >> ifTrue: [ >> Smalltalk globals at: #OneShotArmed put: >> >> 16673 >> <https://pharo.fogbugz.com/f/cases/16673/Senders-of-highlights-wrong-code-se…> >> Senders of... highlights wrong code segment >> >> But I didn't take the "class-comment"-search results into account. >> >> >> >> >> >> Is it a well know bug? >> >> Stef >> >> > >

Oct. 29, 2015

Re: [Pharo-dev] Nautilus not garbage collected ?

by stepharo

Le 29/10/15 06:30, Yuriy Tymchuk a Ã©crit : > Thanks Cyril. This is already solved in the upcoming version of quality assistant. But maybe we can add a test case like that, so in the future this will not repeat? yes excellent idea and thanks for QA! Thanks you cyril too. It is cool to have you around. Stef > > Uko > >> On 29 Oct 2015, at 10:45, Ferlicot D. Cyril <cyril.ferlicot(a)gmail.com> wrote: >> >> Hi, >> >> Yesterday I needed to browse the instances of >> FTPluggableIconListMorphAdaptor and I saw that I had 117 of these. >> So I browsed the instances of Nautilus and I saw I had a lot with no >> window open. >> >> On a new image I ran: >> >> 10 timesRepeat: [ >> |nautilus| >> nautilus := Object browse. >> nautilus close. >> ]. >> >> Then print: >> >> >> 10 timesRepeat: [ Smalltalk garbageCollect ]. >> Nautilus allInstances size. >> >> I get 10. >> >> If I launch the same in Pharo 4 (just replacing Nautilus by >> PackageTreeNautilus) I get 0. >> >> I checked the class that points Nautilus instances: >> >> (Nautilus allInstances flatCollectAsSet: #pointersTo) collect: #class. >> >> And I saw that QANautilusPlugin points to them. >> If I disable QANautilusPlugin the new instances of Nautilus are garbage >> collected. >> >> I am not sure but I think this comes from the SystemAnnoucement. >> >> Should I open a bug entry ? >> >> -- >> >> Cyril Ferlicot >> >> http://www.synectique.eu >> >> 165 Avenue Bretagne >> Lille 59000 France >> > >

Oct. 29, 2015

Re: [Pharo-dev] Nautilus not garbage collected ?

by Yuriy Tymchuk

Thanks Cyril. This is already solved in the upcoming version of quality assistant. But maybe we can add a test case like that, so in the future this will not repeat? Uko > On 29 Oct 2015, at 10:45, Ferlicot D. Cyril <cyril.ferlicot(a)gmail.com> wrote: > > Hi, > > Yesterday I needed to browse the instances of > FTPluggableIconListMorphAdaptor and I saw that I had 117 of these. > So I browsed the instances of Nautilus and I saw I had a lot with no > window open. > > On a new image I ran: > > 10 timesRepeat: [ > |nautilus| > nautilus := Object browse. > nautilus close. > ]. > > Then print: > > > 10 timesRepeat: [ Smalltalk garbageCollect ]. > Nautilus allInstances size. > > I get 10. > > If I launch the same in Pharo 4 (just replacing Nautilus by > PackageTreeNautilus) I get 0. > > I checked the class that points Nautilus instances: > > (Nautilus allInstances flatCollectAsSet: #pointersTo) collect: #class. > > And I saw that QANautilusPlugin points to them. > If I disable QANautilusPlugin the new instances of Nautilus are garbage > collected. > > I am not sure but I think this comes from the SystemAnnoucement. > > Should I open a bug entry ? > > -- > > Cyril Ferlicot > > http://www.synectique.eu > > 165 Avenue Bretagne > Lille 59000 France >

Oct. 29, 2015

Nautilus not garbage collected ?

by Ferlicot D. Cyril

Hi, Yesterday I needed to browse the instances of FTPluggableIconListMorphAdaptor and I saw that I had 117 of these. So I browsed the instances of Nautilus and I saw I had a lot with no window open. On a new image I ran: 10 timesRepeat: [ |nautilus| nautilus := Object browse. nautilus close. ]. Then print: 10 timesRepeat: [ Smalltalk garbageCollect ]. Nautilus allInstances size. I get 10. If I launch the same in Pharo 4 (just replacing Nautilus by PackageTreeNautilus) I get 0. I checked the class that points Nautilus instances: (Nautilus allInstances flatCollectAsSet: #pointersTo) collect: #class. And I saw that QANautilusPlugin points to them. If I disable QANautilusPlugin the new instances of Nautilus are garbage collected. I am not sure but I think this comes from the SystemAnnoucement. Should I open a bug entry ? -- Cyril Ferlicot http://www.synectique.eu 165 Avenue Bretagne Lille 59000 France

Oct. 29, 2015

[pharo-project/pharo-core] f83e5f: 50410

by GitHub

Branch: refs/heads/5.0 Home: https://github.com/pharo-project/pharo-core Commit: f83e5fe78c68ef6c5db1145682de91b1a0a18830 https://github.com/pharo-project/pharo-core/commit/f83e5fe78c68ef6c5db11456… Author: Jenkins Build Server <board(a)pharo-project.org> Date: 2015-10-29 (Thu, 29 Oct 2015) Changed paths: R ScriptLoader50.package/ScriptLoader.class/instance/pharo - scripts/script50409.st A ScriptLoader50.package/ScriptLoader.class/instance/pharo - scripts/script50410.st R ScriptLoader50.package/ScriptLoader.class/instance/pharo - updates/update50409.st A ScriptLoader50.package/ScriptLoader.class/instance/pharo - updates/update50410.st M ScriptLoader50.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st M Spec-Tools.package/MessageBrowser.class/instance/announcements-registration/registerToAnnouncements.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/classModified_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/classRenamed_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/handleClassRenamed_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/handleMethodAdded_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/handleMethodModified_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/handleMethodRemoved_.st M Spec-Tools.package/MessageBrowser.class/instance/announcements/methodAdded_.st M Spec-Tools.package/MessageBrowser.class/instance/announcements/methodModified_.st M Spec-Tools.package/MessageBrowser.class/instance/announcements/methodRemoved_.st A Spec-Tools.package/MessageBrowser.class/instance/announcements/shouldRefreshItem_fromAnnouncement_.st A Spec-Tools.package/MessageBrowser.class/instance/private/intervalOf_inCode_of_.st A Spec-Tools.package/MessageBrowser.class/instance/private/intervalOf_inCommentText_.st M Spec-Tools.package/MessageBrowser.class/instance/text selection/autoSelect_.st A StartupPreferences.package/StartupPreferencesLoader.class/class/menu/startupGeneralPrefererencesFolderMenuOn_.st M StartupPreferences.package/StartupPreferencesLoader.class/class/menu/startupLoaderMenuOn_.st A StartupPreferences.package/StartupPreferencesLoader.class/class/menu/startupPreferencesVersionFolderMenuOn_.st A StartupPreferences.package/StartupPreferencesLoader.class/class/menu/systemStartupMenuOn_.st Log Message: ----------- 50410 16883 Highlight in Senders of... throws a MNU if the search term was found in a class comment https://pharo.fogbugz.com/f/cases/16883 16818 Renaming a Class triggers debugger https://pharo.fogbugz.com/f/cases/16818 16886 Quick access to startup preferences folder https://pharo.fogbugz.com/f/cases/16886 http://files.pharo.org/image/50/50410.zip

Oct. 29, 2015

[pharo-project/pharo-core]

by GitHub

Branch: refs/tags/50410 Home: https://github.com/pharo-project/pharo-core

Oct. 29, 2015

Re: [Pharo-dev] Ubuntu fullscreen problems

by J.F. Rick

Cool. I guess if there is a bug report on squeak then there is no reason to create one on pharo. If you need someone to test a solution on Ubuntu 15.10, let me know. On Wed, Oct 28, 2015 at 11:21 AM Nicolai Hess <nicolaihess(a)gmail.com> wrote: > 2015-10-28 15:35 GMT+01:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com>: > >> I think there must be some incompatibility between Pharo and the host >> window manager/desktop environment. >> >> On Debian/dwm (I am using Ubuntu VM) I can reproduce (1), but not (2). >> >> There is however interesting behavior... if I try to switch desktop while >> in fullscreen it doesn't switch, however once I switch back to >> non-fullscreen, suddenly I jump to the other desktop, as if the keyboard >> shortcut for the switch got buffered somewhere... >> >> Opening a issue in FogzBugz is a good idea so it doesn't get forgotten. >> >> Peter >> >> On Wed, Oct 28, 2015 at 3:15 PM, J.F. Rick <self(a)je77.com> wrote: >> >>> Hi everybody, >>> >>> I just upgraded to Ubuntu 15.10 and am using Unity User Interface. I >>> want Pharo to use the full screen (i.e., above the menu bar and launcher). >>> Inside Pharo, I can choose "Toggle full screen mode". This works to some >>> extent but has some problems: >>> >>> (1) I can no longer switch applications or desktops (i.e., pager). So it >>> stops me doing work where I move between Pharo and a browser. This may be >>> expected functionality. >>> >>> (2) More problematically, when I try to switch out of full screen mode >>> by choosing "Toggle full screen mode", the Pharo interface completely >>> disappears (just a giant gray area) and no mouse or keyboard events seem to >>> register. I have to use command line to kill it. In good news, I can use >>> "save and quit" instead and relaunching it just has it in maximum mode >>> (full screen but launcher and menu bar are visible). I can then toggle full >>> screen mode twice to get back to full screen. >>> >>> I searched for this on FogBugz but found nothing. Is there a workaround? >>> This seems like a VM issue. Is there a newer VM that fixes this? I just >>> used the most recent one on the Pharo 4.0 download page. >>> >>> If nobody has a fix, I'll open a case on FogBugz. This isn't a large >>> problem but it would be nice to have this >>> >> > It is a know problem (with possible fixes) > (http://bugs.squeak.org/view.php?id=7812) > You can not switch back from fullscreen on a linux 64-bit system ( > http://bugs.squeak.org/view.php?id=7833) (maybe not only 32bit issue, but > just the more recent windowmanager) > > > >> >>> Cheers, >>> >>> Jeff >>> >> >> >

Oct. 28, 2015

Re: [Pharo-dev] #50409 regression

by Thierry Goubier

2015-10-28 16:19 GMT+01:00 stepharo <stepharo(a)free.fr>: > I see this is annoying that comments are not polymorphic. > They could be :) - an empty RBMethodNode with a comment attribute, or - a comment: message send with the comment as a parameter. Thierry > > Stef > > Le 28/10/15 11:05, Nicolai Hess a Ã©crit : > > > > 2015-10-28 15:33 GMT+01:00 stepharo <stepharo(a)free.fr>: > >> Hi >> >> I typed 'Take' and source code with it. >> - the progress bar does not update >> >> - more important when I click on the message browser I get a DNU ast >> >> autoSelect: aSelector >> textModel readSelectionBlock: [ :text | >> self selectedMessage >> ifNotNil: [ :msg | >> (self sourceIntervalOf: aSelector in: msg ast) >> ^^^^^^^^^^^^^ >> ifEmpty:[ self searchedString: aSelector asString >> in: text ] ]] >> >> and RGCommentDefinition does not understand ast. >> So what do we do? >> > > It is a know bug ( at least, I know of it :) ) > I changed the way the MessageBrowser selects the search term, > because it always highlights the search term by "text" and not by "code" > ,for example, in a "at:put:" senders MessageBrowser > the first the highlighting started at the first > at: > although this is part of a at:ifAbsent: - message. > > at: #OneShotArmed ifAbsent: [ true ]) > ifTrue: [ > Smalltalk globals at: #OneShotArmed put: > > 16673 > <https://pharo.fogbugz.com/f/cases/16673/Senders-of-highlights-wrong-code-se…> > Senders of... highlights wrong code segment > > But I didn't take the "class-comment"-search results into account. > > > > > > >> >> Is it a well know bug? >> >> Stef >> >> > >

Oct. 28, 2015