- Pharo-dev - lists.pharo.org

[Pharo-project] Fwd: Issue 2551 in pharo: Performance degradation of WeakRegistry in Pharo 1.1
by StÃ©phane Ducasse June 19, 2010

June 19, 2010

Begin forwarded message: > From: Joachim Geidel <joachim.geidel(a)onlinehome.de> > Date: June 19, 2010 5:29:56 PM GMT+02:00 > To: StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> > Subject: Re: Issue 2551 in pharo: Performance degradation of WeakRegistry in Pharo 1.1 > > Hello Stef, > > I just saw that you have integrated Collections-ul.362 to fix the > performance problem with WeakRegistry finalization. > > For Squeak, I had also reported a deadlock > (http://bugs.squeak.org/view.php?id=7546) In a discussion with Levente > Uzonyi on the Squeak-dev list, we found that this was because > WeakKeyDictionary executes the finalizers for gc'd elements in the scope of > the protected block in WeakRegistry>>finalizeValues. As the WeakRegistry has > no control over what the finalizers do, this is a potential source for > deadlocks. This actually happens in JNIPort. The problem was introduced in > Squeak 4.1. > > Levente has fixed the problem in Collections-ul.365.mcz. He has not yet > closed the Squeak issue 7546 yet; I have sent him a reminder. I think it > would be better to integrate Collections-ul.365.mcz. Otherwise there will be > deadlocks in Pharo too, at least with JNIPort. I can work around this, as > JNIPort comes with its own subclass of WeakRegistry, but other users of > WeakRegistry could run into the same problem. OTOH, I don't know what other > changes have been introduced between Collections-ul.362 and > Collections-ul.365, so there may be unwanted changes which are not good for > Pharo. > > Cheers, > Joachim > > ------ Weitergeleitete Nachricht >> Von: "pharo(a)googlecode.com" <pharo(a)googlecode.com> >> Antworten an: "codesite-noreply(a)google.com" <codesite-noreply(a)google.com> >> Datum: Sat, 19 Jun 2010 17:05:59 +0200 >> An: Joachim Geidel <joachim.geidel(a)bluecarat.de> >> Betreff: ***SPAM*** Re: Issue 2551 in pharo: Performance degradation of >> WeakRegistry in Pharo 1.1 >> > >> Updates: >> Status: Fixed >> >> Comment #6 on issue 2551 by stephane.ducasse: Performance degradation of >> WeakRegistry in Pharo 1.1 >> http://code.google.com/p/pharo/issues/detail?id=2551 >> >> (No comment was entered for this change.) >> >> -- >> You received this message because you starred the issue. >> You may adjust your issue notification preferences at: >> https://code.google.com/hosting/settings > > ------ Ende der weitergeleiteten Nachricht > >

1 0

[Pharo-project] [update 1.2] #12010
by StÃ©phane Ducasse June 19, 2010

June 19, 2010

12010 ----- - Issue 2556: AutoStart - Issue 2557: NewTextMorph and CodeMorphs refactorings and tests Fernando apparently one of your test is breaking. can you check? Stef

1 0

[Pharo-project] this is another test
by StÃ©phane Ducasse June 19, 2010

June 19, 2010

1 0

[Pharo-project] this is a test
by StÃ©phane Ducasse June 19, 2010

June 19, 2010

1 0

[Pharo-project] status of the recreateSpecialArray?
by StÃ©phane Ducasse June 18, 2010

June 18, 2010

so what is the status of this discussion? It would be good that we can load alien in 1.1. David what is the status from VMMaker and the latest VMs? http://code.google.com/p/pharo/issues/detail?id=1360 Stef

1 0

[Pharo-project] Mac carbon VM goes to 4.2.5beta1U
by John M McIntosh June 18, 2010

June 18, 2010

In order to wrap up some VM fixes that should be pushed into the Squeak 4.x offering I've compiled up a 4.2.5beta1U VM Someone should run the Sunit and smoke test to ensure the VM is sane. Follow the macintosh link from http://www.squeakvm.org/index.html 4.2.5b1 We update to VMMaker-jcg.182, and we make a new FFI plugin using the latest FFI code base (1.7.0) which includes Josh Gargus changes for FFI ExternalForms to support the OpenCL code. -- =========================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ===========================================================================

3 2

[Pharo-project] ISSUE 2557 : NewTextMorph Refactorings and Tests
by Fernando olivero June 18, 2010

June 18, 2010

SLICE-Issue-2557-NewTextMorphRefactorings-FernandoOlivero.11 I've published to the inbox several major refactoring s to NewTextMorph and CodeMorph. Added a bunch of tests too. The basic idea behind CodeMorph is that an instance of codeMorph holds a target and displays a certain aspect of it. For example the first example displays the method source and the second, only the selectorWithArguments, of the same method. morph := CodeMorph new. morph target: ( CodeMorph >> #styler: ). morph beCode. self assert: morph text asString = 'styler: aStyler styler := aStyler ' . morph beSelectorWithArguments. self assert: morph text asString = 'styler: aStyler'. Lukas, i think now we would be able to easily provide a single line editor for OB, when performing renaming or other selector/arguments or class name refactorings. If you are interested let me now and we can code them together. ( take a look at NewTextMorphTest>>testSelectorWithArguments ). Fernando

2 1

[Pharo-project] MC breakage in 1.1
by Julian Fitzell June 18, 2010

June 18, 2010

It seems that in Pharo 1.1 MC no longer displays the Version window after a failed commit attempt (due to missing authentication for example). As a result, cannot easily resubmit the version after fixing the credentials. In fact, as I look further, it's worse than that. Let's say I had version 1 and changed two methods. After the failed commit the following seems to be the situation: - The ancestor of my working copy is 2 instead of 1 (confirmed with History) - The working copy is dirty - The Changes button shows the same two modified methods that I was trying to save into 2 - The diff between 1 and 2 shows the two modified methods - The diff between 2 and my image (as selected from the History window) shows no changes Eek! Julian

6 15

[Pharo-project] Prof Stef browser
by Schwab,Wilhelm K June 17, 2010

June 17, 2010

For fun, open two Prof Stef browsers and see what happens when you click on the tree in one of them. Then close the second one, and click on the tree in the first. They should act independently. Beyond just a nice system for newbies, this is a potentially important example to get people started building GUIs in Pharo, so the details are important.

2 3

[Pharo-project] Shout colors in 1.1 dev
by Schwab,Wilhelm K June 17, 2010

June 17, 2010

I was trying to figure out why nil was "part of the comment" in a line like self loadMetacelloPackageLatest:'Seaside' version: "'Seaside 2.8'" nil. when I realized that lots of things are colored the same as comments :( One of the most useful things about syntax coloring is to make comments clearly separate from the rest of the code. The defaults could use some tweaking.

1 0