December 2009 - Pharo-dev - lists.pharo.org

[Pharo-project] Email notifications from issue tracker?
by Simon Denier Dec. 4, 2009

Dec. 4, 2009

Hi I wonder why activity from the issue tracker do not get posted on the mailing list? Is it a conscious decision? Because it seems like there are two places for discussion right now: the mailing list and the issues comments. -- Simon

6 6

[Pharo-project] code coverage?
by John McIntosh Dec. 4, 2009

Dec. 4, 2009

Ok as I'm testing this new VM I"m wondering has anyone done a code coverage review of the Pharo image in relationship to the suite of available tests? I wondering how much of the image side code is actually used? Part of the problem here in cross checking the list of internal plugins is that I need test cases to actually exploit the plugin code so that I can confirm it works, even to confirm it gets called. No doubt from the other side it would be interesting to see from a VM platform perspective if the Smalltalk Sunits actually execute a reasonable set of the slang/objective-c/C code. -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

5 9

[Pharo-project] Trouble with using Rome
by Alexandre Bergel Dec. 4, 2009

Dec. 4, 2009

Hi! There is the file RomePlugin.bundle in one folder of the 4.2.1 VM. When I execute "RomeDemo drawCar", I have a rollback saying a primitive has failed. The guilty call is: -=-=-=-=-=-=-=-=-=-=-=-= RomePluginCanvas>>primOpen <primitive: 'primitiveOpen' module: 'RomePlugin'> ^self primitiveFailed -=-=-=-=-=-=-=-=-=-=-=-= The following is displayed in the xterm: Squeak-Rome: canvas oop invalid! Anyone can help? Cheer, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

1 0

Re: [Pharo-project] newinspector?
by Adrian Lienhard Dec. 4, 2009

Dec. 4, 2009

Great, thanks! Adrian On Dec 4, 2009, at 13:07 , Frederic Pluquet wrote: > Hello Adrian, > > I deleted the Elements child in CompiledMethod (it had no sense in > CompiledMethod). > I added also added a literals child to inspect CompiledMethod literals > easily. > > FrÃ©d > ---- > FrÃ©dÃ©ric Pluquet > UniversitÃ© Libre de Bruxelles (ULB) > Assistant > http://www.ulb.ac.be/di/fpluquet > > > On Thu, Dec 3, 2009 at 9:44 PM, Adrian Lienhard <adi(a)netstyle.ch> > wrote: > >> Hi Frederic, >> >> I just run into a problem: clicking on Elements of a NewInspector >> on a >> CompiledMethod gives an error: 'CompiledMethods may only be created >> with >> newMethod:header:' >> >> Shall I create an issue in the Pharo tracker? >> >> Adrian >> >> >> On Nov 26, 2009, at 19:40 , Adrian Lienhard wrote: >> >> Just yesterday, I asked Frederic to remove the class and method >>> entries that add a lot of the noise and are most often not needed. >>> He >>> committed a new version with these changes. >>> >>> But from Stef's analysis it seems there is a bit more needed to make >>> this inspector fly... And a really good inspector is key! >>> >>> Another item to add to Stef's list: what is the point of the item >>> "Keys"? I can see the keys of the dictionary already by opening >>> "Elements". So this seems superfluous. >>> >>> I wonder why inspecting a dictionary could not be reduced to the >>> information that is actually of interest, namely the keys and their >>> values. So when opening a dictionary, the subtree would directly >>> display the keys in the left pane (no array, tally, Keys, Elements). >>> If a key is selected, its value is printed in the right pane. >>> Expanding a key would dive into its associated value. Actually, I >>> think what I described here is more or less the behavior of the old >>> explore tool. >>> >>> Cheers, >>> Adrian >>> >>> >>> On Nov 26, 2009, at 18:39 , StÃ©phane Ducasse wrote: >>> >>> Hi guys >>>> >>>> again I did a demo of smalltalk and the newInspecotr got in my way. >>>> I tried to analyze what did not work: >>>> - apparently the refresh does not work. Does change in >>>> instance >>>> variables get refreshed in the inspector? >>>> >>>> I try to understand what get in my way in general. >>>> - when an object has few instance variable I think that it >>>> is lost >>>> in the sea of trees. >>>> >>>> - I think that the |> class is not useful at the top level. >>>> May be just having Class: and the class Name is enough no >>>> need for >>>> the |> >>>> May be the Class : ByteString should appear after the >>>> Elements: >>>> >>>> - the |> methods to be useful should be sorted by category >>>> >>>> in the screenshot >>>> Why do I see the instance variable of the dictionary >>>> (array and >>>> tally) I find that confusing? >>>> Because I get elements and at the same time the instance >>>> variables. >>>> >>>> I do not understand why some lines are bold and other blue >>>> >>>> When we click on connectInfo I see on the right pane >>>> - size : 5 >>>> [#host] : a SocketAddress (size: 24) >>>> [#loginMethod] : #zork >>>> [#password] : nil >>>> [#port] : 110 >>>> [#user] : 'stephane.ducasse' >>>> But I cannot modify anything. >>>> >>>> What is the point to have Elements expanded in the left >>>> pane when >>>> I >>>> get the contents by clicking on the instance >>>> variable and getting the contents on the right. >>>> >>>> Then Keys should probably close to Elements. >>>> >>>> I do not see the point to have |> for individual method >>>> I do not understand why we see Elements and ByteCode >>>> I get a DNU when I click on Elements of a method BTW >>>> >>>> >>>> <Screen shot 2009-11-26 at 6.14.31 >>>> PM.png>_______________________________________________ >>>> Pharo-project mailing list >>>> Pharo-project(a)lists.gforge.inria.fr >>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>> >>> >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>> >> >>

1 0

[Pharo-project] allowUnderscoreAssignment is not functional
by Alain Plantec Dec. 4, 2009

Dec. 4, 2009

Hi all, maybe I've missed something but whatever value is set to this preference, the result is that underscore is always allowed. Note that it is used by the scanner: Scanner>>xUnderscore Preferences allowUnderscoreAssignment ifFalse:[^self xIllegal]. self step. tokenType := #leftArrow. ^token := #':=' The question behind is do we keep this preference ? Cheers Alain

3 2

[Pharo-project] cmdDotEnabled
by Alain Plantec Dec. 4, 2009

Dec. 4, 2009

the same as for allowUnderscoreAssignment, cmd-dot is always functional whatever allowUnderscoreAssignment preference value. do we keep the preference ? Alain

4 5

[Pharo-project] [update 1.1] #11070
by Marcus Denker Dec. 4, 2009

Dec. 4, 2009

11070 ----- Issue 1539: remove class ProjectLoading Issue 1540: remove more unsent messages... Issue 1537: vmSatisticsReportString averages always show 0 at first decimal place Issue 1529: Date and time classes should not rely on collections arithmetic

2 1

Dec. 4, 2009

Begin forwarded message: > From: Lukas Renggli <renggli(a)me.com> > Date: December 4, 2009 8:18:48 AM GMT+01:00 > To: StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> > Subject: Re: what is the status of this one? _ in selector > > Thanks for letting me know about the bugs that I can fix, because I don't really follow the incoming bugs. > > I fixed the one with the small input box: > > http://code.google.com/p/pharo/issues/detail?id=1375 > > That one I marked as invalid: > > http://code.google.com/p/pharo/issues/detail?id=1355 > >> What is the status of this one? _ in selector > > > I can provide an easy fix of the Pharo compiler. Another possibility would be to provide a fix with a preference to toggle between the two modes, with the new preference system this is now much simpler. > > The problem of the preference is that the system could be in an inconsistent state and not allow a recompilation anymore. > > So what kind of change do you want? > > Lukas > > -- > Lukas Renggli > http://www.lukas-renggli.ch >

1 0

Dec. 3, 2009

Hi guys I continued to code on PackgeInfo killing and I changed completely the internal structure of an RPackage. All the tests are green now. Now I should read what john and don did in RBSelectorEnvironment because this is really similar. After I should deal with - class renaming - hooking into the system (ugly) - how to make sure that * are sucked in and sucked out. Some people told me that they would like to help and you are welcome. The code is in PharoTaskForces/RPackage Stef

1 0

Dec. 3, 2009

11069 ----- Issue 1541: Settings integration part3 *added a menu entry in the world menu for the new setting browser. * added double-click handling in MorphTreeMorph * SystemSettingBrowser, double clicking on a row open a class browser on the method which declares the selected setting * do not update a SystemSettingBrowser if it is not viewed by an opened window (prevents updating of a closed but not yet garbaged window) * improved gradient background rendering (radial direction) * added System-Support-SystemSettings, Network-SystemSettings, Polymorph-SystemSettings and Monticello-SystemSettings packages * migrate all preferences from the old Browsing * migrate existing http proxy preferences * migrate existing media preferences * add settings for network user name, password, blab email, author fullname * deprecation of some old preference methods Thanks a lot alain!!!!!!

3 8