- Pharo-dev - lists.pharo.org

[Pharo-project] A russian-chars input problem
by Dennis Schetinin April 29, 2009

April 29, 2009

I use pharo0.1-10268web09.04.1.image. Unable to type in some russian capital letters (specifically 'Ð', 'Ð' , 'Ð§' and 'Ð¢') which are Shift+ J, K, X and N accordingly (in russian layout ofcourse). Shift+J seems to be just ignored. Shift+K opens World menu. Shift+X gives a pair of apostophes (''). Shift+N gives a pair of quotes (""). -- Dennis Schetinin

3 4

[Pharo-project] Slow Image removes window decoration
by stephan＠stack.nl April 29, 2009

April 29, 2009

An hour later, the pharo application goes full-screen on the second monitor, losing the window border, including title bar and icons. The updating continues, and an hour later finishes. I've tried interrupting the update process, but those interrupts only arrive after the update finishes. After saving and restarting, I now have a Pharo image without a window border. Reproducing the issue might be, uhhm, slow. Stephan

1 0

[Pharo-project] isqueak aka iphone vm and closures
by John M McIntosh April 29, 2009

April 29, 2009

I've build a testing vm that enables closures for the iPhone. I've not shipped yet. I have not yet build a patch set for the touch event logic, I'll wait for the general event handling stuff to settle first. However I was able to start the pharo web image with all the latest updates from yesterday on the iphone and see that seaside/pier worked as expected. -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

2 1

[Pharo-project] SUnit bug ?
by Mariano Martinez Peck April 28, 2009

April 28, 2009

Hi folks! Today, at work, a friend ask me about a method in TestResource because he thought it was a bug. We look at it, and for us, it could be a bug. So, I ask to you. See TestResource class side isAvailable ------------------------ isAvailable ^self current notNil and: [self current isAvailable] ------------------------ the first part (self current notNil) ALWAYS return true because of the implementation of self current: ------------------------ current current isNil ifTrue: [current := self new]. ^current ------------------------- The second part too. Because the implementation of self available (object side method): ------------------------- isAvailable "override to provide information on the readiness of the resource" ^true ------------------------- Ok, you can override and change this method in subclasses, however, the first part is ALWAYS true. Is this correct ? or isAvailable should be: isAvailable ^ self current isAvailable Thanks! Mariano

1 0

[Pharo-project] about libnui and licensing
by Sebastien Metrot April 28, 2009

April 28, 2009

Hi, I'm the author of libnui which has been discussed here two days ago. (i saw this list archive in the referer stats for libnui.net and was curious ;-) ). Some people seemed interested in nui but had concerns about it's licensing. I summarized our choice of license in this wiki page: http://redmine.libnui.net/wiki/libnui/Licensing . We're not particularly happy with the GPL but we found no real other option that permitted to fulfill our requirements listed on the page. As others have noted C++ is a very ugly language and recreating a useable object model have been quite a chore. I'm curious about how we could change nui to make it more friendly to other languages such as smalltalk, knowing that C++ has always been a tad harder to interoperate than, for example, C. Thanks in advance, SÃ©bastien

2 1

[Pharo-project] Any south american in this list?
by Alexandre Bergel April 28, 2009

April 28, 2009

I am pretty sure that yes. Would you mind to ping me? I will visit Argentina (La Plata / Bueno Ares) at the end of July. A Pharo party may be organized? Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

6 7

[Pharo-project] Cmd-D does not work anymore?
by Alexandre Bergel April 28, 2009

April 28, 2009

I found this pretty handy. I can have a look at it if nobody yet did so. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

3 5

[Pharo-project] Slow image
by astares＠gmx.de April 28, 2009

April 28, 2009

Anyone with a win32 system able to confirm that the new image slows heavily down as descibed in [1] using the procedure in [2]? After a first rough test I think it happens after the "Delay" fixes from update 10287/10288 Thx T. [1] http://lists.gforge.inria.fr/pipermail/pharo-project/2009-April/007910.html [2] http://lists.gforge.inria.fr/pipermail/pharo-project/2009-April/007934.html -- Neu: GMX FreeDSL Komplettanschluss mit DSL 6.000 Flatrate + Telefonanschluss fÃ¼r nur 17,95 Euro/mtl.!* http://dslspecial.gmx.de/freedsl-surfflat/?ac=OM.AD.PD003K11308T4569a

1 0

[Pharo-project] Suspending process fix
by Igor Stasenko April 28, 2009

April 28, 2009

Hello, i trying to provide a fix to the issue http://bugs.squeak.org/view.php?id=6822 what is needed to make following code work w/o error: |sema proc | sema := Semaphore new. proc := [ sema wait. self error: 'should not be there' ] fork. Processor yield. proc suspend. proc resume. The problem here, is that once you issue a #suspend on a process it puts nil to its myList ivar, ignoring the fact that it could be waiting on semaphore. This leads to disconnecting the process from semaphore, and any subsequent 'sema signal' will be simply lost: suspend | oldList | <primitive: 88> "primitive will fail, if receiver is not active process, which is just our case" myList ifNil:[^nil]. oldList := myList. myList := nil. oldList remove: self ifAbsent:[]. ^oldList the possible fix would be: - add an instance variable to Process, say 'semaphore'. - fix the #suspend to following: suspend | oldList | <primitive: 88> "primitive will fail, if receiver is not active process, which is just our case" myList ifNil:[^nil]. oldList := myList. myList := nil. oldList remove: self ifAbsent:[]. oldList isKindOf: Semaphore ifTrue: [ semaphore := oldList ]. ^oldList - then, in the #resume resume suspendedContext ifNil: [^ self primitiveFailed]. semaphore ifNotNil: [ semaphore resumeWaiting: self. semaphore := nil. ^ self. ] ^ self primitiveResume and Semaphore>>resumeWaiting: aProcess excessSignals>0 ifTrue: [excessSignals := excessSignals-1. aProcess primitiveResume. "resume immediately" ] ifFalse: [self addLastLink: aProcess] ---------- An attached file is the changeset which makes following test to work w/o failure: |sema proc bool | bool := true. sema := Semaphore new. proc := [ sema wait. bool ifTrue: [self error: 'should not be there']. Transcript show: 'semaphore signaled' ] fork. Processor yield. proc newSuspend. proc newResume. self assert: (sema last == proc). bool := false. sema signal. P.S. i didn't checked how this fix will affect other things around processes, like mentioned by Mike: Note that Process>>signalException: (as of in Squeak 3.8, it may be fixed in later versions?) relies on this broken behaviour. -- Best regards, Igor Stasenko AKA sig.

1 0

[Pharo-project] Random failures with Linux VM
by Michael Rueger April 27, 2009

April 27, 2009

Hi all, while trying to update a 10281 on the closure (exupery) Linux VM I'm running into random failures like DNUs, failed network access etc. I'm running ubuntu in VirtualBox. Has anybody seen anything similar? Maybe I should try to build a "normal" Linux VM, just to check... Michael

5 9