September 2012 - Pharo-dev - lists.pharo.org

[Pharo-project] Max CPU usage when #drawOn: is overridden
by S Krish Sept. 4, 2012

Sept. 4, 2012

If we override #drawOn: and implement custom logic to draw.. every few milliseconds as it gets called the CPU usage shoots to available max.. Run four instances of the same class that overrides and does some moderately complex draw routine of say : CircleMorph with border set, few lines, color rendering.. I will send in the .st file later .. that can be tested against.. Is this expected or what are the ways to avoid this CPU overhead... NativeBoost animation, far more responsive.. also goes up very similarly. but little more tolerable... VGTigerMorph... more so than FlakesDemo...

1 1

[Pharo-project] [NativeBoost] Question: zero-based or one-based indexes?
by Igor Stasenko Sept. 3, 2012

Sept. 3, 2012

Hi, there i added a bunch of accessors to ByteArray and NBExternalAddress which having uniform look: nb<type>At: nb<type>At: put: where types are: {U}Int<8/16/32/64> or Float<32/64> the problem is that indexes in ByteArray are 1-based, but in NBExternalAddress is 0-based. And there's already different places in code which following these rules. But i feel like this should be uniform (so same piece of code can be used for either bytearrays or external addresses), to avoid confusion and mistakes. i just not sure which one to leave: 1-based or 0-based ? >From one side, all collections in smalltalk is 1-based.. but from other side accessing memory at 1-based offset looks also unnatural. What you think? -- Best regards, Igor Stasenko.

10 19

[Pharo-project] A useful set of Smalltalk-like methods for Objective-C
by Oscar E A CallaÃº Sept. 3, 2012

Sept. 3, 2012

Hi all, I just release a useful set of Smalltalk-like methods (eg. collect: or match: ) for Objective-C classes NSArray, NSMutableArray and NSString. I just extends those classes using categories and blocks. I made this, because when programming on Objective-C, I really miss them. You can find those methods on: https://github.com/ocallau/ObjCSmalltalkExtensions They are in an iOS project, but you can use them in OSX too. Of course, they come with a set of unit tests (on GHUnit). I hope these methods can be useful for you, as they are for me. Cheers. Oscar

2 2

[Pharo-project] What happens with Warning when running in headless image (2.0)?
by Dale Henrichs Sept. 3, 2012

Sept. 3, 2012

I'm running a headless pharo2.0 from a script and the vm disappears without a trace (no pharodebug.log info etc.) ... when I run the script in a gui, I see that a warning dialog box is popping up, so perhaps this is the source of the issue? ... or perhaps it's something else? Dale

4 10

[Pharo-project] [NativeBoost] 1.4 Bleeeding edge release
by Igor Stasenko Sept. 3, 2012

Sept. 3, 2012

Hi, i was in a mood of making some overhauls.. so as result i refurbished a type system a bit.. removed all bogus C-like class names , e.g. NBUnsignedIntType ... and replaced them with nice and uniform names, like NBUInt32, NBInt16 etc. (thanks Marten Feldtmann for an idea) A uniform type aliases for these types also added: int8 NBInt8 uint8 NBUInt8 int16 NBInt16 uint16 NBUInt16 int32 NBInt32 uint32 NBUInt32 int64 NBInt64 uint64 NBUInt64 the rest of C integer types are actually aliases of the above. And some new features: - finally spent time for implementing support of 64-bit integer types. (strangely over 2 years nobody needed them ;) - added a bunch of uniform accessors to ByteArray and NBExternalAddress. But please avoid heavily depending on using them right now, before we decide whether they should use uniform indexing, or stay 1-based index for byte arrays, but 0-based indexes for external addresses. And if we decide to have them uniform, should they use 1-based or 0-based indexing. May be force with you. P.S. The configuration is freshly-baked, i didn't even checked if it loads cleanly or not.. so, i don't recommend updating it in image with your precious code :) Special instruction to load NB for Stef: Gofer it squeaksource: 'NativeBoost'; package: 'ConfigurationOfNativeBoost'; load. (ConfigurationOfNativeBoost project version: '1.4' ) load P.P.S. i would really like some help with testing what i did, including writing some more tests to cover type system more broadly (now since type names follow certain naming rule, many different types can be checked without much coding) -- Best regards, Igor Stasenko.

3 3

[Pharo-project] Internationalization in Pharo - Updates?
by Gisela Decuzzi Sept. 2, 2012

Sept. 2, 2012

Hello, I want something of i18n for Pharo 1.4. I want to add english support for Loop. Searching I found 2 ways rigth now: - GetText package (ported from eToys), the problem here is that seems to be not working in Pharo 1.4 - the translate message The last updates I could find were years ago and I would like to know if there is any work related (in progress or done) or if someone with experience could point me the path to follow and avoid reinventing the wheel. I would be pleaced to hear any comment, even if this is some pending task that needs to be done or a link to a thread I coluldn't find. Thanks in advance!

3 3

[Pharo-project] interesting consonance
by Igor Stasenko Sept. 2, 2012

Sept. 2, 2012

C side <-> Seaside. /me hacking NB and thinking about C side ;) -- Best regards, Igor Stasenko.

1 0

[Pharo-project] [Nautilus] Someone needs to learn how to work with weak references :) + [Enh] patch to object explorer
by Igor Stasenko Sept. 2, 2012

Sept. 2, 2012

my work image grown to 80Mb in size.. and i start growing suspicious about myself for forming a resource leak somewhere.. But as it appears, it is because of Nautilus. During my experiments with ObjC wrapper for NativeBoost, i used nautilus for showing more than 4000 Objective-C classes in browser.. Since all those objects are actually wrappers, they generated each time when i open a new browser window.. to explore an obj-c universe and to test bugs.. So i ended up with: ObjCClassWrapper allInstances size 31885 Now, about why cache doesn't cleaning up automatically: The only reference coming to particular instance of ObjC wrapper is from a ClassesIconCache class var, which is a weak key dictionary: {my class wrapper} -> [IconicButtonStateHolder] so, a {wrapper} is held weakly here, of course, while [its Icon] strongly, and so far so good. But the problem is that , [IconicButtonStateHolder] having following in it's arguments ivar: an Array(an ObjCClassWrapper), which is [my class wrapper], held strongly. of course, there is no way how it can disappear from system.. After doing "AbstractNautilusUI resetIconCaches" my image shrunk from 83Mb down to 43Mb, But i found that i still having a lot of ObjCClassWrapper hanging around and wonder why.. And this time, because one of the wrapper classes were held in HistoryCollection via Nautilus 'RecentClasses' classvar. I think, history should not keep a live object references.. but instead use symbolic names to refer to the target, because apparently you cannot go and browse the method/class if you deleted it from a system. So, after nilling out that class var, my image lost another 10Mb, down to 33Mb.. and Ben, you owe me a ... (o shit i don't drink beer ) oh ok.. you owe me 2 bug fixes in Nautilus :) P.S. i modified a pointer explorer to show only strong references to object(s), because weak ones are completely useless for fighting with resource leak problems and only add noise. With this change a lists produced by pointer explorer are much shorter and going straight to the point. So, i attaching .cs to the mail.. in case if someone find it useful. -- Best regards, Igor Stasenko.

4 9

[Pharo-project] [ANN] Nagare - a new flexible logger for Smalltalk
by Masashi UMEZAWA Sept. 2, 2012

Sept. 2, 2012

Hello all, I have recently developed a new logging library called Nagare. http://code.google.com/p/nagare-logger/ Nagare is designed to be simple and scalable. Contrary to other logging frameworks which provide rich set of log output methods (file, socket, DB, etc.), Nagare just connects to fluentd (http://fluentd.org) And Fluentd does the various log processing jobs. Features: - Simple log interfaces with reliable backend - Semi-structured logging (Not only String, you can store structured records in log) - Flexible - you can easily customise log-tags, log policy, etc. - Portable (runs on Squeak, Pharo, VisualWorks) Enjoy! -- [:masashi | ^umezawa]

2 2

[Pharo-project] An idea about visual feedback in Athens
by Igor Stasenko Sept. 2, 2012

Sept. 2, 2012

so, the problem is that all you see on screen consists from shapes, which are drawn by Athens. Now, since drawings can be quite complex and include various transformations, it will be hard to track them manually to use for mouse (or fat finger) pointer hits. A simplest example: i wanna draw a rectangle, rotated by 45 degrees. and i wanna change it's color when mouse is over it. If you look at the situation from global coordinate system perspective, in order to find whether some morph contains given point or not, you have to iterate over all it's parent morphs (because they also can have coordinate system transformations). But in a local coordinate system, things remain extremely simple: to test whether a point is inside of rectangle (most morphs usually is) is a piece of cake, isnt? I think i found an interesting approach for a coordinate system feedback, (so it can be used for 'hit tests' as well as many other things) So, lets consider a simple draw method: Morph>>drawOnAthensCanvas: aCanvas aCanvas setPaint: Color red; drawShape: self bounds now if we add a single line here: self registerShapeForEvents: self bounds canvas: aCanvas. Things getting very interesting: what this method should do, is to simply capture enough state from current canvas (like transformation matrix, clipping and the shape , of course), so later it can be used to test whether some point (like mouse cursor position) is over it or not. Why in drawing method , you might ask? Where else, i would answer.. A parent morphs branch can impose own clipping and transformation, which you can easily observe during drawing, but it becomes tedious once you leave outside of it: to get same amount of information you'll have to manually go over all morph's hierarchy and apply same transformation/clipping calculations once you get to the morph under the question. The drawing method responsible for visual representation of morph on display media, which means that it actually defines the shape(s) with which users will interact. So, by drawing a bunch of morphs on desktop using such approach, we can automatically capture all the geometry (as well as its relative hierarchy), which needs to be tested for mouse hits (or receiving a mouse events in general). And the hit test , which we need to perform every mouse move, boils down to simply iterating over that hierarchical list and testing the current hand position against each element. The morphs, which are not interested in receiving a mouse events (like hidden or parent morphs of various kinds), apparently can simply avoid publishing anything into that list, instead of implementing numerous #handlesMouseDown/#handlesMouseOver etcc... and like that saving cycles for being tested every time you touching the mouse, because some of their submorphs are actually interested in receiving such events. And of course a most interesting aspect of this approach is handling a complex geometry for morphs. Because with rectangles it is fairly simple to imagine a visual feedback system which based solely on morph's bounds (like the one which we currently using), but for morphs with complex geometry, things can become quite tedious. -- Best regards, Igor Stasenko.

2 6