August 2016 - Pharo-dev - lists.pharo.org

when Cmd-P can just display a string in the debugger

by stepharo

Hi Seriously I do not get why printing in the debugger cannot just print the text without putting these ^%&*^*&()**(&%^*( double quote around. Why do we change such basic functionality and in particular we do not propose ANY alternative. If you need a pop up and do not care of manipulating text this is ok but WHY the default behavior has to be changed. I hate so much that the GTTool ALWAYS want to IMPOSE a flow. Your &^%*^(& flow is not mine. Should I go and hack to get back the default behavior. WHY do I have to do that? Seriously. Especially since the solution was super simple: introduce a new binding living nicely close to PrintIt but no "we have to change the default one with something that may fuck your flow but who cares because my flow is better." Why do we get systematically this message? Why such changes are damaging TDD practices without any notices? May be I should use another Smalltalk at the end. Stef

Aug. 7, 2016

Re: [Pharo-dev] [pharo-project/pharo-core] f57843: 60163

by stepharo

Le 5/8/16 Ã 05:28, Martin Dias a Ã©crit : > Hi Stef, good, I take note. I improved filters in the recent releases > and will continue improving. Thanks Martin I have problems with the menu item names :) To me redo looks strange when I get a crash I want to import the changes not redo them. Then I do not get what undo means (yes I know this is probably revert) May be some pharoers have better names than mine. Stef > > Martin > > On Wed, Aug 3, 2016 at 5:30 PM, stepharo <stepharo(a)free.fr > <mailto:stepharo@free.fr>> wrote: > > Thanks Martin. > > We missed yesterday: show only the latest version of every > entities. :) > > > Setf > > > Le 3/8/16 Ã 11:23, Marcus Denker a Ã©crit : > > On 03 Aug 2016, at 11:07, GitHub <noreply(a)github.com > <mailto:noreply@github.com>> wrote: > > > Log Message: > ----------- > 60163 > 18831 Integrate new Epicea version > https://pharo.fogbugz.com/f/cases/18831 > <https://pharo.fogbugz.com/f/cases/18831> > > ChangeLog for this: > > - Case 18813: Implement redo and undo of protocol addition and > removal. > > - Case 18612: disable drag&drop until TreeModel supports it > correctly. > > - Case 18384: Redo and undo: show on any selection (not only > when all selected entries are code changes). > > - Enhancement: Improve lost changes detection; show only the > events actually lost. > > - Enhancement: Pass on filters. > > - Clean up: Do not show the "absent entry" item anymore since > it was not useful at all. > > - Clean up: Remove unused "commit tags" in log browser. > > - Clean up: Remove unused tests. They were good idea, we can > re-introduce them but working and covering more cases. > > - Clean up: Rename #displayWidget to #asMorph in > EpLogBrowserItem''s hierarchy. It is more meaningful. > > - Workaround: Hide "Open in Sorter" menu action since it is > useless until Treemodel drag&drop is fixed. > > > > 18768 Inlined method const could be implemented by metalinks > https://pharo.fogbugz.com/f/cases/18768 > <https://pharo.fogbugz.com/f/cases/18768> > > 18835 Update RBParser-Nodes class comments > https://pharo.fogbugz.com/f/cases/18835 > <https://pharo.fogbugz.com/f/cases/18835> > > http://files.pharo.org/image/60/60163.zip > <http://files.pharo.org/image/60/60163.zip> > > > > > > >

Aug. 7, 2016

Re: [Pharo-dev] Refactoring HelpTopic (needs review)

by Nicolai Hess

2016-07-28 16:18 GMT+02:00 stepharo <stepharo(a)free.fr>: > I will! > > I love to read nice comments > > Stef > > I updated the slice and added some more example (much more is needed but I think it is better than having no doc :) ). I changed some help topics on refactoring support in nautilus to directly used the refactoring classes comments (no need to maintain two docs), but this is not applicable for all refactorings, for example "generate superclass" and "insert superclass" needs its own description, as there is only one RBAddClassRefactoring comment. please help with reviewing, I am not good at writing this kind of text. The refactoring engine is a great and powerfull piece of code and it would be great to have a proper documentation "in-image". I have seen some bugs, bad code or wrong bug fixes by contributors not fully understanding the framework (me too :) ). just load the slice and open HelpBrowser openOn: RefactoringHelp. thanks nicolai > Le 28/7/16 Ã 00:38, Nicolai Hess a Ã©crit : > > Hi, > > I recently added some class comments to the Refactoring-Package. > (RBBrowserEnvironments, RBRefactoring, RBNamespace). > > Now I added some more HelpTopics describing the refactoring framework, > some background information, RBPatternParser and Metavariables, some > examples and put the existing class comments into the HelpBook. > > see fogbugz issue 18829 > <https://pharo.fogbugz.com/f/cases/18829/Add-a-RefactoringFramework-System-H…> > Add a RefactoringFramework System HelpTopic > you can load the slice and open the HelpBook with > > HelpBrowser openOn: RefactoringHelp. > > Would be nice if someone can help with proofreading and of course, amy > feedback is welcome. > > nicolai > > >

Aug. 6, 2016

Re: [Pharo-dev] Moving Monkey quality checks to Renraku infrastructure

by stepharo

Well, so far I did not yet investigate all that part. I'm actually experimenting on a new monkey implementation that has the following objectives: > - easy to configure and run locally > - faster. It should be able to run build validations in parallel > (e.g., in my prototype tests run by 4 parallel pharo images are run in > 2 minutes) > - it should enforce the same process used for issue validation and > integration > - integrated with the bootstrap :) This is super cool: it will change the way we work and make our life a lot simpler. Stef PS: I'm happy to work with motivated, visionary smart and efficient guys.

Aug. 6, 2016

Re: [Pharo-dev] Moving Monkey quality checks to Renraku infrastructure

by stepharo

Le 4/8/16 Ã 18:39, Yuriy Tymchuk a Ã©crit : > Hi, > > At the moment I am moving Pharo quality tools to Renraku model. This is a quality model that Iâve been working on and that was used so far by QualityAssistant. > > At the moment Iâm stuck while trying to do changes in Monkey, as it is really hard to understand how quality checks are made there. @Guille maybe you can advise something. > > In the old model Rules were both checking code and storing the entities that violate them. In Renraku Rules are responsible only for checking, then for each violation they produce a critique object that is a mapping between the rule and the entity that violates the rule. Also critiques can provide plenty of additional information such as suggestion on how to fix the issue. This is good that you improve this part. The critics browser was done by someone that was told to do it and not someone that believe that it is important :) So it was a good first step. Now this is good to revisit LintRules. They already helped us a lot. Rules are really important for Pharo (and you know it) so this is supercool that you push that. Thanks a lot Yuriy. > At the moment we can skip the critiques altogether, just run the rules and store the classes and method that violate them. But at the moment I cannot how Monkey in implemented. I.e. where do the rules come from and what output should be provided. > > Cheers. > Uko >

Aug. 6, 2016

Re: [Pharo-dev] Pharo website broken on HTTPS

by stepharo

Thanks for the notice. This is strange since we did not change anything. So clearly holiday period of some admin layer. Stef Le 5/8/16 Ã 19:13, Andres Valloud a Ã©crit : > Also FYI the Pharo Consortium's website at > http://consortium.pharo.org/ displays a Proxy Error. > > On 8/5/16 7:35 , Esteban A. Maringolo wrote: >> Hi all, >> >> The Pharo website doesn't work properly if you access it vÃa HTTPS >> (and Chrome), apparently some resources (JS and CSS) are explicitly >> linked to HTTP, so if you access it by HTTPS those resource are >> blocked by the client. >> >> See attached screenshot. >> >> Regards! >> >> Esteban A. Maringolo >> > >

Aug. 6, 2016

another segfault from Pharo5.0 vm (freshly downloaded today) Ubuntu 14.04

by Dale Henrichs

Attached crash dump file ... This one occurred after I'd had an image open for several hours interesting that similar to the other crashes I've seen recently FreeTypeFace seems to be implicated: Smalltalk stack dump: 0xff7bc02c I [] in FreeTypeFace(FT2Handle)>pvtDestroyHandle 0xcdeae10: a(n) FreeTypeFace 0xff7bc04c M BlockClosure>ensure: 0x9edc9b0: a(n) BlockClosure 0xff7bc078 I [] in Mutex>critical: 0xadc7960: a(n) Mutex 0xff7bc098 M [] in Semaphore>critical: 0xba42d50: a(n) Semaphore 0xff7bc0b8 M BlockClosure>ensure: 0x9edcab8: a(n) BlockClosure 0xff7bc0d8 M Semaphore>critical: 0xba42d50: a(n) Semaphore 0xff7bc100 I Mutex>critical: 0xadc7960: a(n) Mutex 0xff7bc124 I FreeTypeFace(FT2Handle)>pvtDestroyHandle 0xcdeae10: a(n) FreeTypeFace 0xff7bc13c M FreeTypeFace(FT2Handle)>finalize 0xcdeae10: a(n) FreeTypeFace 0xff7bc154 M ByteSymbol(Symbol)>value: 0xa3ab850: a(n) ByteSymbol 0xff7bc178 M ObjectFinalizerCollection(OrderedCollection)>do: 0xaa62580: a(n) ObjectFinalizerCollection 0xff7bc19c I ObjectFinalizerCollection>finalize 0xaa62580: a(n) ObjectFinalizerCollection 0xff7bc1c0 I WeakFinalizerItem>finalizeValues 0xb8cae80: a(n) WeakFinalizerItem 0xff7bc1dc M [] in WeakRegistry>finalizeValues 0xa6dd690: a(n) WeakRegistry 0xff7bc1f4 M BlockClosure>on:do: 0x9edc8a8: a(n) BlockClosure 0xff7bc214 M BlockClosure>on:fork: 0x9edc8a8: a(n) BlockClosure 0xff7bc234 M [] in WeakRegistry>finalizeValues 0xa6dd690: a(n) WeakRegistry 0xff7bc258 M OrderedCollection>do: 0x9edc450: a(n) OrderedCollection 0xff7bc280 M WeakRegistry>finalizeValues 0xa6dd690: a(n) WeakRegistry 0xff7bc29c M [] in WeakArray class>finalizationProcess 0xa5cefc0: a(n) WeakArray class 0xff7bc2b4 M BlockClosure>on:do: 0x9edc398: a(n) BlockClosure 0xff7bc2d4 M BlockClosure>on:fork: 0x9edc398: a(n) BlockClosure 0xff7bc2f4 M [] in WeakArray class>finalizationProcess 0xa5cefc0: a(n) WeakArray class 0xff7bc318 M WeakArray(SequenceableCollection)>do: 0xa381328: a(n) WeakArray 0xff7bc33c I [] in WeakArray class>finalizationProcess 0xa5cefc0: a(n) WeakArray class 0xff7bc35c M [] in Semaphore>critical: 0xcc147b0: a(n) Semaphore 0xff7bc37c M BlockClosure>ensure: 0x9edb760: a(n) BlockClosure 0xff7bc39c M Semaphore>critical: 0xcc147b0: a(n) Semaphore 0xff7bc3c0 I WeakArray class>finalizationProcess 0xa5cefc0: a(n) WeakArray class 0xbd19fe8 s [] in WeakArray class>restartFinalizationProcess 0xce10050 s [] in BlockClosure>newProcess Looks like pvtDestroyHandle went a little overboard:) Dale

Aug. 5, 2016

Re: [Pharo-dev] Pharo website broken on HTTPS

by Andres Valloud

Also FYI the Pharo Consortium's website at http://consortium.pharo.org/ displays a Proxy Error. On 8/5/16 7:35 , Esteban A. Maringolo wrote: > Hi all, > > The Pharo website doesn't work properly if you access it vÃa HTTPS > (and Chrome), apparently some resources (JS and CSS) are explicitly > linked to HTTP, so if you access it by HTTPS those resource are > blocked by the client. > > See attached screenshot. > > Regards! > > Esteban A. Maringolo >

Aug. 5, 2016

GUI for switching to gitfiletree metadata-less mode

by Peter UhnÃ¡k

Hi, since I am slowly switching all my repositories to metadata-less, it would be cool if there was some GUI for it. The best place would be when right clicking on the repository in the monticello browser, unfortunately that menu is hardcoded. An alternative is to add another button the repository detail, i.e.: MCFileTreeGitRepositoryInspector>>buttonSpecs repository readonly ifTrue: [ ^ super buttonSpecs ] ifFalse: [ ^ #(#('Pull' gitPull 'git pull and refresh') #('Push' gitPush 'git push' gitNeedPush)) , super buttonSpecs , (repository hasMetadata ifTrue: [ #(#('Remove Metadata' removeMetadata 'switch to metadata-less mode')) ] ifFalse: [ #() ]) ] MCFileTreeGitRepositoryInspector>>removeMetadata "Switch the repository to a metadata-less mode" | properties | properties := STON fromString: repository directory / '.filetree'. properties at: 'Metadata' put: 'false'. (repository directory / '.filetree') ensureDelete; writeStreamDo: [ :stream | stream nextPutAll: (STON toJsonStringPretty: properties) ]. repository gitCommand: #('commit' '-m' 'Switched repository to metadata-less mode.' '--' '.filetree') in: repository directory. repository flushCache. self refresh Not the prettiest, but it seems to work. Btw I've written a short post for this https://www.peteruhnak.com/blog/2016/08/05/switch-to-metadata-less-gitfilet… Peter

Aug. 5, 2016

Pharo website broken on HTTPS

by Esteban A. Maringolo

Hi all, The Pharo website doesn't work properly if you access it vÃa HTTPS (and Chrome), apparently some resources (JS and CSS) are explicitly linked to HTTP, so if you access it by HTTPS those resource are blocked by the client. See attached screenshot. Regards! Esteban A. Maringolo

Aug. 5, 2016