- Pharo-dev - lists.pharo.org

[Pharo-project] Pinesoft enhancements
by Bill Schwab June 28, 2008

June 28, 2008

Gary, It appears to have loaded. I did the MCPackageLoader hack described on the wiki. It appeared to have frozen one time, and succeeded on a second attempt. As always, Thanks! Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bschwab(a)anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029

1 0

[Pharo-project] EToys removal
by Gary Chambers June 27, 2008

June 27, 2008

Just to say that my investigation is almost complete. Having completed most of our work on the various Pinesoft Sprints I should have a bit more time for it... Obviously there is some overlap with MVC/Nebraska/ScriptLoader removal (and a bit of UI Enhancements modifications) though. Just a concern, forgiving a quote from Alain's email, "Stephane seems to prefer that I work on Etoy removing". I'm wondering the best way forward with this now. Regards, Gary.

3 5

[Pharo-project] Proposal for external issues
by Norbert Hartl June 27, 2008

June 27, 2008

Every now and then someone tries to improve the way to go by introducing new processes. Most of time nobody likes them and they fail :) We have several types of changes that we want to integrate in pharo: ï»¿- enhancements to the platform - fixes for reported defects - fixes from another squeak release (e.g. 3.10) - fixes from other bug databases like mantis The type of an issue you open on google is pretty near to that. Just think about it as a source where the change has its origin. For this purpose I created a Type-Mantis in the tracker to assign to tickets from mantis we like to integrate. It might be a step more to do when opening a ticket but the outcome is that we track a list of things already done. if you look at http://code.google.com/p/pharo/issues/list?can=1&q=label%3AType-Mantis you can see what I mean. I think this way it is easier to track things. What do you think? Norbert

3 5

[Pharo-project] Squeak bugs, Pharo bugs
by Matthew Fulmer June 27, 2008

June 27, 2008

Hi. I am wondering what the process is of filing bugs against Pharo/Squeak, notably ones already in Squeak mantis. This bug against Networking is longstanding, and has a fix: http://bugs.squeak.org/view.php?id=6768 Should I try to get it into Pharo? Is there a list of Mantis issues that have been folded into Pharo? Is there a process to determine which mantis bugs apply to Pharo? I ask because Pharo is currently, de facto, the only work being done on Squeak core, and I don't see a good reason to really push a separate effort and call it 3.11. So, how are you using squeak mantis in Pharo? -- Matthew Fulmer -- http://mtfulmer.wordpress.com/

5 8

[Pharo-project] what happened to the "load code updates" button in the left flap?
by Oscar Nierstrasz June 27, 2008

June 27, 2008

Hi guys, Haven't had time to look at Pharo for a while. I grabbed the latest version (Pharo0.1-10042-2008-06-24.zip) and there is no flap on the left hand side. Is there a short cut to load the code updates? The Pharo-dev version (pharo-0.1-10036dev08.06.20.zip) still has the flap. Did I miss something? - on

5 5

[Pharo-project] Fwd: [squeak-dev] Units package
by StÃ©phane Ducasse June 27, 2008

June 27, 2008

> > we should harvest the Units....latest version > > Begin forwarded message: > >> From: "Marcus Denker" <denker(a)iam.unibe.ch> >> Date: June 24, 2008 2:11:45 PM CEDT >> To: "BalÃ¡zs KÃ³si" <rebmekop(a)gmail.com> >> Cc: The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org >> > >> Subject: Re: [squeak-dev] Units package >> Reply-To: The general-purpose Squeak developers list <squeak-dev(a)lists.squeakfoundation.org >> > >> >> Hello, >> >> the bug seems to be fixed in the last version. I harvested the test >> to make sure (this is >> in Units-md.16 on SqueakSource) >> >> The versions starting with -14 contain many enhancements (and the >> bugfix). This was contributed >> by Pinesoft (http://www.pinesoft.co.uk/) >> >> Maybe the universe version could be updated? >> >> Marcus >> >> -- >> Marcus Denker -- denker(a)iam.unibe.ch >> http://www.iam.unibe.ch/~denker >> >> >> On Thu, Jun 19, 2008 at 4:56 PM, BalÃ¡zs KÃ³si <rebmekop(a)gmail.com> >> wrote: >> > UnitValue >> hash >> > ^value hash bitXor: unit hash >> > >> > and >> > >> > testEquality >> > >> > self deny: (60 units: #seconds) hash = (60 units: #meters) >> hash. >> > self deny: (60 units: #seconds) = (60 units: #meters). >> > self deny: (60 units: #seconds) hash = (1 units: #minutes) >> hash. >> > self assert: (60 units: #seconds) = (1 units: #minutes). >> >> Much better :) >> >> Cheers, BalÃ¡zs >> >> ps: cc-d to Marcus Denker, because he seems to be maintainer of the >> package. >> >> >> >> > > >

4 4

[Pharo-project] [ANN] new update
by StÃ©phane Ducasse June 26, 2008

June 26, 2008

here is a couple of fixed. BTW yesterday I issued another image. Stef Couple of changes to make more tests work plus fix of finalization logic: - changed WeakRegistry to use WeakIdentityKeyDictionary (http://bugs.squeak.org/view.php?id=6347 ) - FinalizerTest reverted to real test case - added missing class variables to DigitalSigntureAlgorithm - Fix issue #13: TraitsTests pollutes ProtoObject subclasses (Mantis 7090) - make BitmapStreamTest clean up the file bitmapStreamTest.ref.extSeg ------------------------------------------------------ update10043 MVC cleaning step 3 ------------------------------------------------------ update10044 cleaning unimplemented calls ------------------------------------------------------ update10045 Some more cleaning made by Norbert ------------------------------------------------------ update10046 KernelTests-Numbers -- Nan and hex tests broken Collection methods #select: and #collect: have no test MC fixes Remove testAllMethodsWithSourceString

3 8

[Pharo-project] ReleaseTest>>testUnimplementedCalls
by Norbert Hartl June 26, 2008

June 26, 2008

I thought once more about this test. I'm reluctant to have this test removed. On the other hand is Adrien right. He states that a lot of test classes are doing calls to methods they created while the test runs. One solution would be to change the test to use perform: to invoke those methods. But that would making tests a bit tedious. So this is not my preferred solution. I tweaked SystemNavigation>>allUnimplementedNonPrimitiveCalls to allUnimplementedNonPrimitiveCallsX "Answer an Array of each message that is sent by an expression in a method but is not implemented by any object in the system." | all meth dict | dict := Dictionary new. all := self systemNavigation allImplementedMessages. self systemNavigation allBehaviorsDo: [:cl | cl selectorsDo: [:sel | meth := cl compiledMethodAt: sel. meth primitive = 0 ifTrue: [ meth messages do: [:m | (all includes: m) ifFalse: [ ((dict at: cl ifAbsentPut: [ Dictionary new ]) at: sel ifAbsentPut: [ OrderedCollection new]) add: m ] ] ] ] ]. ^ dict So the test would be like (SystemNavigation default allUnimplementedNonPrimitiveCallsX associationsSelect: [:a| (a key allSuperclasses includes: TestCase) not ]) isEmpty Would this or something similar be useful or is this just a lot of noise without much of a benefit? Norbert

3 2

June 25, 2008

Begin forwarded message: > From: "Samuel MORELLO" <samuel.morello(a)ouvanous.com> > Date: June 25, 2008 2:09:06 PM CEDT > To: "StÃ©phane Ducasse" <stephane.ducasse(a)inria.fr> > > http://www.metacafe.com/watch/yt-oNpH5tx4Pu8/lighthouses_and_waves_during_v… > > -- > Samuel > +33 (0)6 20 65 22 99 >

1 0

June 25, 2008

Hi all I got some funny unpublished packages and I'm a bit confused. So could you check what I did. I was harvesting MVC removal + norbert cleaning SLICE- UnimplementedCallsFix1-NorbertHartl.1 I have to find some time (argh) to improve the process and the tools. because this is far too much manual. Stef

3 21