- Pharo-dev - lists.pharo.org

Re: [Pharo-project] Pharo Kernel break...
by Marcus Denker March 6, 2012

March 6, 2012

On Mar 4, 2012, at 10:32 AM, Pavel Krivanek wrote: > Hi, > > I shortly looked at it. To reproduce the key problem without need to > start shrinking process we may use this expression. > > Behavior removeCategory: '*Deprecated14' Strange... a workaroung would do to just unload it with Monticello instead. We should merge the shrinking of #cleanUpForProduction with the one of the Kernel... Marcus -- Marcus Denker -- http://marcusdenker.de

2 2

[Pharo-project] Nautilus collisions in icons
by Mariano Martinez Peck March 6, 2012

March 6, 2012

Hi ben. I found that method's icons are not only one, and there can be colisions. For example, I have test methods which came from a trait. So I have the trait icon rather than the test. And I can imagine some other colision as well ... So...what can we do? can we have a list of icons? I mean..have as much icons as requiered? (instead of having colisions and let only one) If we cannot have multiple...then I would let the test icon in the case of traits for example....because I don't care where the method come from...the same way I don't care if it comes from a superclass, I don't care it comes from a trait. But I do care that if it is a test, I could do cmd+t and see the result :) what do you think? Cheers -- Mariano http://marianopeck.wordpress.com

2 1

[Pharo-project] NativeBoost VM's
by Jan van de Sandt March 5, 2012

March 5, 2012

Hello, On the Pharo Jenkins server [1] there is a recent build of a VM with NativeBoost support for the Mac. But the Unix version is six months old and there is no Windows version. Is this just a Jenkins issue and is NativeBoost working on Unix and Windows? Or are these platforms not working yet? Jan. [1] https://ci.lille.inria.fr/pharo/view/NativeBoost/

3 3

Re: [Pharo-project] Showing the setUp method in Nautilus
by Alexandre Bergel March 5, 2012

March 5, 2012

Voila, I produced version 4.1 Please check. I did not use a plugin since the commands are directly implemented on NautilusUI, I therefore guess that a plugin is not the way to define commands. Also, something that I would find very useful, is to open a new Nautilus directly on the Group tab. I often see myself switching to the Group tab. Am I the only one? Cheers, Alexandre On 5 Mar 2012, at 13:40, Benjamin wrote: > > On Mar 5, 2012, at 12:52 PM, Alexandre Bergel wrote: > >> Hi! >> >> I worked on a new method item action to easily show the setUp method of a testCase. This helps reduce the amount of clicks and context switching when one wants to know what the setUp method is about while writing a test method. > > Cool > >> I think it is a good candidate for being included in Nautilus. >> Ben, if you agree with this, I can to commit in the Nautilus rep and produce a new version of the configuration. How does that sound? > > Commit, and I will have a look ;) > I though you started also a NautilusPlugin related to that, am I wrong ? > >> >> Question: how to associate a shortcut to an action? > > Anywhere on class side something like: > > buildRefactoringMethodShortcutsOn: aBuilder > <keymap> > > (aBuilder shortcut: #rename) > category: #NautilusMethodShortcuts > default: $r command > do: [:target || scroll | > scroll := target methodWidget vScrollValue. > target refactor renameMethodFor: target selectedMethod. > target methodWidget vScrollValue: scroll] > > The categories are: > #NautilusMethodShortcuts > #NautilusProtocolShortcuts > #NautilusClassShortcuts > #NautilusPackageShortcuts > > and here target is the NautilusUi instance which get the keyStroke. > >> >> Side comment: it would be a good move to adopt a class hierarchy a la OB to define commands. Defining all the commands in NautilusUI does not scale and simply does not work for non trivial command. > > For me, it generates to much classes. And now that Guillermo did a great work for KeyMappings, I prefer to use it :) > > > Ben > >> >> Cheers, >> Alexandre >> -- >> _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: >> Alexandre Bergel http://www.bergel.eu >> ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. >> >> >> >> <ShowSetupInNautilus.1.cs> > -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

4 6

[Pharo-project] MVP
by Schwab,Wilhelm K March 5, 2012

March 5, 2012

Hello all, I have been to busy to read today, but I like seeing any discussion of MVP. I am **slowly** working on a framework for it, but would be glad to be beaten to it. Default models are key; they must be replaceable, but (sub)triads should work "out of the box." Models can include value holders, value adapters, converters, and contexts, depending on how they are wired. I have something of a start on the adapters and converters, which I can try to package and make available. If nothing else, they will have value in seeding discussion if not be directly usable. More to come. Bill

5 21

[Pharo-project] Cog VM binaries as per VMMaker.oscog-eem.152/r2538.
by Eliot Miranda March 5, 2012

March 5, 2012

Cog VM binaries as per VMMaker.oscog-eem.152/r2538<http://www.mirandabanda.org/files/Cog/VM/VM.r2538/> . Fix tricky context state bug that can cause crashes in the GC. Contextpart>runUntilErrorOrReturnFrom: may pop the last argument. Hence if a frame is married in this state its spouse context won't have its last argument slot initialized. If later the frame is divorced and the arguments are not updated the last argument slot can be left with this uninitialized slot and ... bang. The solution is to update arguments as well as stack contents when divorcing. This is good also because it removes the VM's assumption that method arguments are read-only, and that's only enforced by the Smalltalk bytecode compiler, not by the bytecode. -- best, Eliot

3 3

Re: [Pharo-project] MVP
by Marcus Denker March 5, 2012

March 5, 2012

On Mar 5, 2012, at 7:25 PM, Schwab,Wilhelm K wrote: > At least in the 1.4 seaside image, I do not get color highlighting the browser - debugger and workspaces do show > coloring. I just grabbed the latest seaside image to verify. > then the build scripts of that image are wrong. A "seaside" image should not load anything related to browsers, workspaces or debuggers. Marcus -- Marcus Denker -- http://marcusdenker.de

3 2

Re: [Pharo-project] MVP
by Marcus Denker March 5, 2012

March 5, 2012

On Mar 5, 2012, at 6:45 PM, Schwab,Wilhelm K wrote: > I can use 1.4 to look at the code. Once shout works in the browsers, I might make the switch anyway. > In Pharo 1.4 as downloaded here: https://ci.lille.inria.fr/pharo/job/Pharo%201.4/lastSuccessfulBuild/artifac… Shout is working in the Browsers. Marcus -- Marcus Denker -- http://marcusdenker.de

1 0

Re: [Pharo-project] small thoughts... system browser to show super methods & wide debugger
by Marcus Denker March 5, 2012

March 5, 2012

On Mar 5, 2012, at 11:33 AM, Ben Coman wrote: > > Just sharing some small thoughts as they cross my mind on the chance the are of any interest. Not necessarily looking for any response or action. > > 1. In the System Browser when viewing a method, this might include an expandable list of overridden methods. Beneath the code of the current method being viewed each 'super' method would be listed with an expandable '+' symbol. I have attached a mock up [1] showing all expanded but the default might be the collapsed state. > > 2. The debugger lists instance and temporary variables to be inspected, but not every object that 'passes through' the method being executed is held in a variable. More often messages are chained together such that to inspect the object returned from the middle of this chain you need to select that bit of the code and choose to inspect it again. However there may be side-effects of executing that a second time so you are not guaranteed to end up looking at the object actually encountered by the execution of the method. > What would be interesting would be if in certain circumstances every object returned by every message in a method was held to be inspected. This might take the form of hovering or right-clicking on any message an choosing something like 'inspect context'. The circumstances this might be turned on for could be: > a. globally > b. ongoing execution from within debugger > c. certain methods / classes / categories. > > 3. With screens getting much bigger, perhaps the debugger could have an option to show code from previous calls back to a certain level. I have attached a mock up [2]. Even though this image shows using a lot of vertical space, this would be scrolled. Ultimately it would be horizontal space required as the code steps in, but todays wide screens match this even better than those a few years ago. Note the yellow highlighted line in the top pane which can position how far back you look. > I like the debugger idea! Especially combined with mouse-over infos (show the value if var in a pop-up when the mouse it over it), this should allow for a very quick way to understand code... Marcus -- Marcus Denker -- http://marcusdenker.de

2 1

[Pharo-project] [update 1.4] #14378
by Marcus Denker March 5, 2012

March 5, 2012

14378 ----- Issue 5436: Move DialogItemsChooser to Finder (its only user) http://code.google.com/p/pharo/issues/detail?id=5436 Issue 5435: clean up unusedBlocks tests http://code.google.com/p/pharo/issues/detail?id=5435 -- Marcus Denker -- http://marcusdenker.de

1 0