April 2010 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] [squeak-dev] Context reflection regression?

by Lukas Renggli

AFAIK, the values are copied on send to the arguments/temps of thisContext (1...4), they are not kept in the sending stack frame. Lukas On 11 April 2010 11:32, Igor Stasenko <siguctua(a)gmail.com> wrote: > I just found that > > a: a b: b c: c > Â ^ thisContext sender stackPtr > > answers 0 Â :( > > while, in order to call the above method , a caller, obviously needs > to push 4 values on stack: > Â - push receiver > Â - push a > Â - push b > Â - push c > Â - send #a:b:c: > > so, a method's caller context stack _should_ contain at least 4 elements. > > So, either i wrong, or VM doesn't tracking the context's sp value > correctly anymore. > > -- > Best regards, > Igor Stasenko AKA sig. > > -- Lukas Renggli www.lukas-renggli.ch

April 11, 2010

[Pharo-project] Context reflection regression?

by Igor Stasenko

I just found that a: a b: b c: c ^ thisContext sender stackPtr answers 0 :( while, in order to call the above method , a caller, obviously needs to push 4 values on stack: - push receiver - push a - push b - push c - send #a:b:c: so, a method's caller context stack _should_ contain at least 4 elements. So, either i wrong, or VM doesn't tracking the context's sp value correctly anymore. -- Best regards, Igor Stasenko AKA sig.

April 11, 2010

[Pharo-project] Graphics-MichaelRueger.168....backported normalized transform

by stephane ducasse

Hi apparently mike posted a while ago transform changes. There were unnoticed because buried over a lot of other changes. So mike should I integrate these changes. Even when I use the bug entry I tend to forget my own changes so without bug entry item. the chances are even higher. Name: Graphics-MichaelRueger.168 Author: MichaelRueger Time: 17 December 2009, 12:02:55 pm UUID: 9a3cca4d-1a2d-5d4b-84c6-b6d3286b98e0 Ancestors: Graphics-MichaelRueger.167 backported normalized transform

April 11, 2010

[Pharo-project] [update 1.1] #11311

by StÃ©phane Ducasse

11311 ----- - Issue 1842: Make the System -> Software update a preference for 1.1. Thanks alain. - Issue 991: Polymorph text entry dialog is ugly . Thanks alain - Issue 2287: Menu Selection Color. Thanks gary.

April 11, 2010

Re: [Pharo-project] [squeak-dev] Re: SqueakSource is unusable

by Igor Stasenko

On 11 April 2010 00:18, stephane ducasse <stephane.ducasse(a)gmail.com> wrote: > thanks. My thanks too. It is just GREAT improvement! Keep moving guys! :) -- Best regards, Igor Stasenko AKA sig.

April 11, 2010

Re: [Pharo-project] [squeak-dev] Re: SqueakSource is unusable

by jaayer

---- On Sat, 10 Apr 2010 13:29:33 -0700 Bert Freudenberg <bert(a)freudenbergs.de> wrote ---- >On 10.04.2010, at 21:19, stephane ducasse wrote: >> >> >> On Apr 10, 2010, at 7:16 PM, Bert Freudenberg wrote: >> >>> Thanks. >>> >>> I just merged the stuff I developed lately >> >> which is? > >Most importantly, projects can have parent projects now (only if a site-wide setting is enabled). This is to support "inbox" style use - when mailing out diffs, the diff based is searched in a project's parent project too. > >Also, you can enable developers to move versions, not just admins. This is a per-project setting, again to enable developers to clean out an "inbox" repository. There can also be a default target project for moves. > >Since SSProject grew too many instance variables, I added a properties dictionary just like we have for versions and repositories. Refactoring might be in order but I just copied and pasted. Sorry ;) > >There are a few minor changes, see the version history :) > >- Bert - Bert, something else I would really like to see is sticky sessions. It is annoying to continually discover I've been logged out and my session has expired every time I should happen to neglect a SS page open in a browser tab for a little too long. This forces me not only to log back in but also to re-navigate to whatever page I was viewing, as starting a new session takes you back to the homepage. This can be remedied with AJAX, specifically a Prototype or JQuery periodical on an empty <div/> that invokes an empty #refresh callback on the server side every 30 seconds or so. This keeps the session alive as long as a SS page is open. > > >_______________________________________________ >Pharo-project mailing list >Pharo-project(a)lists.gforge.inria.fr >http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 11, 2010

Re: [Pharo-project] Bye bye pharo on the iPhone

by Lawson English

Igor Stasenko wrote: > 2010/4/10 John M McIntosh <johnmci(a)smalltalkconsulting.com>: > >> On 2010-04-10, at 9:08 AM, Stefan Marr wrote: >> >> >>> There are rumors, that this change is motived by technical reasons related to multitasking. >>> I could imagine some nice tricks related to the efforts Apple is putting into LLVM, to actually have a 'smart' C/C++ runtime system which allows to assess what kind of activity profile an app is going to exhibit. >>> This is already hard enough with C, prohibiting any VM technology seems to be a reasonable step, if they are actually going to employ any analysis techniques to get their multitasking stuff 'right'. >>> >>> But this is pure speculation. >>> >>> In the light of Steve Job's remark: "We just shipped it on Saturday, and we rested on Sunday." everything is possible, even that he is just going... >>> >>> >>>>> http://www.macrumors.com/2010/04/09/fallout-from-apples-exclusion-of-flash-… >>>>> >>> The primary reason for the change, say sources familiar with Apple's plans, is to support sophisticated new multitasking APIs in iPhone 4.0. The system will now be evaluating apps as they run in order to implement smart multitasking. It can't do this if apps are running within a runtime or are cross compiled with a foreign structure that doesn't behave identically to a native C/C++/Obj-C app. >>> >>> "[The operating system] can't swap out resources, it can't pause some threads while allowing others to run, it can't selectively notify, etc. Apple needs full access to a properly-compiled app to do the pull off the tricks they are with this new OS," wrote one reader under the name Ktappe. >>> << >>> >> Nonsense. >> >> An hour with some unix internals book and reading a bit about suspend/resume, and reflect on what happens when you sleep your unix based laptop shows there is no magic involved, just a bit of change to how Processes are managed. >> > > +1.. this is a bullshit. > Instead of solving the problem, they locking-down their platform. > > Its like saying "we're going to build an aircrafts with 4 wings, and > from this moment, all two-winged planes should stop being used > worldwide". > > Its just a way of making sure that all iPhone/iPad/Mac development is still done on Macs, IMHO. 10+ years ago, Apple promised developers a way to program Mac OS X apps for Windows. With the advent of QuickTime X, based on Cocoa libs, I've been speculating that Apple was planning on leveraging those libraries as a distribution of Mac OS X frameworks to Windows that 3rd party developers could use. It doesn't seem a total stretch that if Apple does this, they want to make sure that iPhone/iPad apps can only be developed on Mac and by extension, Mac OS X applications, even for Windows, will only be developed on the Mac as well. One IDE to rule them all: Mac OS X, iphone/ipad/iTouch, and now (MAYBE) Windows... Lawson

April 10, 2010

Re: [Pharo-project] SqueakSource is unusable

by jaayer

---- On Sat, 10 Apr 2010 02:33:01 -0700 Lukas Renggli <renggli(a)gmail.com> wrote ---- >> or if not that, then nearly so. > >Thank you for the friendly subject, I almost marked it as spam. Well, perhaps the title was a tad harsh, but I feel the issues addressed in my post were legitimate. I posted it to the list because there are multiple versions of SS on SS, and it is not at all clear who one should contact with issues. >I've integrated you changes, but I had to change many things because >your styles broke the layout of the two columns and the position of >the separation bar between them. If you add this: <div style="clear:both"/> within the .contents <div> but after the .contentsInner <div>, then the the separator will always extend to the bottom, even on pages like the .mcz package details view where the navigation is longer than the contents. (You could also create a .clear class with that property and put the div in it.) This is a hack, but a well-known, widely-employed one. >Lukas > >-- >Lukas Renggli >www.lukas-renggli.ch > >_______________________________________________ >Pharo-project mailing list >Pharo-project(a)lists.gforge.inria.fr >http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

April 10, 2010

Re: [Pharo-project] Initial experience coming from Squeak

by Sean P. DeNigris

Lukas Renggli wrote: > > I've removed the pre-selection and it does not work (or I > don't understand it). > Yes, you're right (it was working for me because I had to click there once to deselect it, duh!). So maybe leaving it deselected and having the user double-click is the best compromise - no savings in steps, but more natural. Sean -- View this message in context: http://n4.nabble.com/Initial-experience-coming-from-Squeak-tp1799044p183567… Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

April 10, 2010

Re: [Pharo-project] Initial experience coming from Squeak

by Lukas Renggli

>> I don't see why not pre-selecting the template helps? I've quickly >> changed that, but clicking on it doesn't select it. > > If the template is not pre-selected, single clicking off the template (in > the pane below the bottom of the template) selects it. As I said, I've removed the pre-selection and it does not work (or I don't understand it). Lukas -- Lukas Renggli www.lukas-renggli.ch

April 10, 2010