- Pharo-dev - lists.pharo.org

Re: [Pharo-project] [Pharo-users] invalid utf8 input detected
by Sven Van Caekenberghe July 27, 2011

July 27, 2011

On 27 Jul 2011, at 05:49, Jon Hancock wrote: > Thanks Sven. I'm using the latest Cog build on the latest pharo 1.3 build on linux with ubuntu system fonts loaded and am using in pharo the same font as my gEdit text editor which displays correctly. > Anyone know why this error occurs or why the fonts show up as question marks. I am trying to consider using pharo for some Chinese projects and need to ensure all string throughout pharo respect UTF8 well. Is the smalltalk code also stored in UTF8 form? > Also, anyone have a trick for capturing the full stack trace from a walkback or debugger as a text stream so I can copy into emails? I think one of the regular Linux users might be better placed to answer, as this relates to UX issues. Internally, Pharo uses Unicode. UTF8 is one of a list of possible encodings (and the most used one for non ASCII). What happens when you paste from the system clipboard to Pharo is some kind of import, propably with conversions going on. This necessarily depends on the VM and OS. On Mac OS X this seems to 'just work' as expected. In your case is seems to fail. Anyone better placed to help Jon ? Sven

1 0

[Pharo-project] [update 1.4] #14061
by Igor Stasenko July 27, 2011

July 27, 2011

- Issue 4555: Fix SUnit to prevent forked process from escaping to debugger http://code.google.com/p/pharo/issues/detail?id=4555 -- Best regards, Igor Stasenko AKA sig.

1 0

[Pharo-project] TestCase>>#should:notTakeMoreThan: behaves not well in headless mode
by Igor Stasenko July 27, 2011

July 27, 2011

It forks a process for evaluating a block you passing. And it works quite well if there's no exception during block evaluation.. but if there is, then it signals the TestFailure exception, not in a process who waits test to be finished, but in a process which forked. And then since there is no exception handler, it using default action, which just to open a debugger. In headless mode it leads to immediate quit from image and logging error report. Like following: THERE_BE_DRAGONS_HERE MessageNotUnderstood: NonInteractiveUIManager>>openDebuggerOn:context:label:contents:fullView: 26 July 2011 6:35:08 pm VM: unix - i686 - linux-gnu - Croquet Closure Cog VM [CoInterpreter VMMaker-oscog-IgorStasenko.108] Image: Pharo1.4a [Latest update: #14025] NonInteractiveUIManager(Object)>>doesNotUnderstand: #openDebuggerOn:context:label:contents:fullView: Receiver: a NonInteractiveUIManager Arguments and temporary variables: aMessage: openDebuggerOn: a Process in nil context: MethodContextTest(TestCase)...etc... exception: MessageNotUnderstood: NonInteractiveUIManager>>openDebuggerOn:contex...etc... resumeValue: nil Receiver's instance variables: uiManager: a MorphicUIManager doNotQuitOnRestart: false Debugger class>>openOn:context:label:contents:fullView: Receiver: Debugger Arguments and temporary variables: process: a Process in nil context: MethodContextTest(TestCase)>>signalFailure: title: 'TestFailure: Assertion failed' contentsStringOrNil: nil bool: false fullView: false Receiver's instance variables: superclass: CodeHolder methodDict: a MethodDictionary(#addOptionalButtonsTo:at:plus:->(Debugger>>#addO...etc... format: 168 instanceVariables: #('interruptedProcess' 'interruptedController' 'contextStack...etc... organization: ('*Polymorph-Widgets' addOptionalButtonsTo:at:plus: buildMorphicN...etc... subclasses: nil name: #Debugger classPool: a Dictionary(#AlwaysOpenFullDebugger->false #ContextStackKeystrokes-...etc... sharedPools: nil environment: a SystemDictionary(lots of globals) category: #'Tools-Debugger' traitComposition: {} localSelectors: nil [:tool | ^ tool openOn: aProcess context: aContext label: aString contents: contents fullView: aBool] in ToolRegistry>>debug:context:label:contents:fullView: Receiver: a ToolRegistry Arguments and temporary variables: aProcess: Debugger aContext: a Process in nil aString: MethodContextTest(TestCase)>>signalFailure: contents: 'TestFailure: Assertion failed' aBool: nil tool: false Receiver's instance variables: tools: an IdentityDictionary(#basicInspector->BasicInspector #browser->[self de...etc... inspectorsMap: an IdentityDictionary(#CompiledMethod->#CompiledMethodInspector ...etc... ToolRegistry>>using:do: Receiver: a ToolRegistry Arguments and temporary variables: aToolName: #debugger aBlock: [:tool | ^ tool openOn: aProcess context: aContext label: aString...etc... tool: Debugger Receiver's instance variables: tools: an IdentityDictionary(#basicInspector->BasicInspector #browser->[self de...etc... inspectorsMap: an IdentityDictionary(#CompiledMethod->#CompiledMethodInspector ...etc... ToolRegistry>>debug:context:label:contents:fullView: Receiver: a ToolRegistry Arguments and temporary variables: aProcess: a Process in nil aContext: MethodContextTest(TestCase)>>signalFailure: aString: 'TestFailure: Assertion failed' contents: nil aBool: false Receiver's instance variables: tools: an IdentityDictionary(#basicInspector->BasicInspector #browser->[self de...etc... inspectorsMap: an IdentityDictionary(#CompiledMethod->#CompiledMethodInspector ...etc... Process>>debug:title:full: Receiver: a Process in nil Arguments and temporary variables: context: MethodContextTest(TestCase)>>signalFailure: title: 'TestFailure: Assertion failed' bool: false topCtxt: Process>>debug:title:full: Receiver's instance variables: nextLink: nil suspendedContext: nil priority: 40 myList: nil threadId: nil errorHandler: nil name: 'Process to evaluate should: notTakeMoreThanMilliseconds:' env: nil Process>>debug:title: Receiver: a Process in nil Arguments and temporary variables: context: MethodContextTest(TestCase)>>signalFailure: title: 'TestFailure: Assertion failed' Receiver's instance variables: nextLink: nil suspendedContext: nil priority: 40 myList: nil threadId: nil errorHandler: nil name: 'Process to evaluate should: notTakeMoreThanMilliseconds:' env: nil TestFailure>>defaultAction Receiver: TestFailure: Assertion failed Arguments and temporary variables: Receiver's instance variables: messageText: 'Assertion failed' tag: nil signaler: MethodContextTest>>#testClosureRestart signalContext: TestFailure(Exception)>>signal handlerContext: nil outerContext: nil .... This happens when you run a MethodContextTest>>#testClosureRestart on Cog VM. Why it fails is another story :) But we need to fix the tester somehow, in order to continue running rest of tests instead of quitting to OS. The problem is that "aBlock value" which signals the exception should be wrapped by exception handler. And then #should:notTakeMoreThan: should report that exception to the main process which runs the test. So, a failure would be either due to timeout or due to failing running the code without exceptions. Any suggestions? -- Best regards, Igor Stasenko AKA sig.

2 3

Re: [Pharo-project] Downloading 1.4
by Marcus Denker July 26, 2011

July 26, 2011

On Jul 26, 2011, at 7:53 PM, Sean P. DeNigris wrote: > > StÃ©phane Ducasse wrote: >> >> https://ci.lille.inria.fr/pharo/job/ >> > > Thanks. It doesn't seem to launch with the Mac Carbon 4.2.5 VM (the only one > with all the plugins built :( ), although it did launch with the Stack > 4.0.0. > > I created Issue 4553. Is it really realistic to support 20 differents VMs? Who is going to test? Marcus -- Marcus Denker Do it together: http://taskfor.us

5 5

[Pharo-project] problem with text editor when editing code and get a proposal to change it
by StÃ©phane Ducasse July 26, 2011

July 26, 2011

Hi guys In 1.4 I get an annoying regression, when the system propose me a change my text gets mangled or not updated. Am I the only one to get that? It would be good to fix that. Stef

4 4

Re: [Pharo-project] Downloading 1.4
by Marcus Denker July 26, 2011

July 26, 2011

On Jul 26, 2011, at 8:34 PM, Igor Stasenko wrote: > On 26 July 2011 21:23, Sean P. DeNigris <sean(a)clipperadams.com> wrote: >> >> Igor Stasenko wrote: >>> >>> and the question remains same. :) >>> As you can see the Carbon VMs which are built by Hudson having 4.0.0 >>> version. >>> So, what is VM you using? >>> >> >> It's the one at http://www.squeakvm.org/mac/release/Squeak%204.2.5beta1U.zip >> >> Is that what you're asking? >> > > Yes. > So, you using an old released VM instead of development version. > Should work. Unless before you run your image with Cog and saved it.. > Then some older VMs refusing to open it. > Yes, and as I don't use an old VM, the image that Jenkins starts with is saved in the new format. And no, I will *not* change that. I changed that for 1.3, but 1.4 is 1.4. We need to move on. > Jenkins server (who builds 1.4 images now) are using quite old unix > VM. So there should be no problems with image format. > (or maybe we already using stack VM for building 1.4 images, which > using new image format?) > The unix vm on the server can read images saved with Cog. Marcus -- Marcus Denker Do it together: http://taskfor.us

2 1

Re: [Pharo-project] Downloading 1.4
by Marcus Denker July 26, 2011

July 26, 2011

On Jul 26, 2011, at 7:58 PM, Sean P. DeNigris wrote: > > Marcus Denker-4 wrote: >> >> Is it really realistic to support 20 differents VMs? Who is going to test? >> > > I'm sure that once the new VMs stabilize, it will not be an issue, but given > that the Stack/Cog VMs do not have all the plugins, what do you suggest for > right now? We should fix the VMs. We can *not* support to run on all old VMs. Marcus -- Marcus Denker Do it together: http://taskfor.us

3 8

[Pharo-project] Downloading 1.4
by Sean P. DeNigris July 26, 2011

July 26, 2011

How to I download 1.4? I don't see it on the website or ci server. Thanks. Sean -- View this message in context: http://forum.world.st/Downloading-1-4-tp3696454p3696454.html Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

2 4

Re: [Pharo-project] Pharo 1.3 Seaside 'client' build ?
by Marcus Denker July 26, 2011

July 26, 2011

On Jul 25, 2011, at 4:31 PM, Sven Van Caekenberghe wrote: > Hi (Marcus), > > I was looking for a Pharo 1.3 Seaside 'client' build on the CI server(s) but couldn't find it. > > It used to be available for 1.2.x (this seems to be gone as well). > > I think that for good reciprocal relations with Seaside, it would be very good to have a tested build of the current main Pharo images. It make a good testcase as well. And it saves quite some work, loading the Seaside config is pretty slow. > > I remember reading something about some clients being removed from the CI's, but I didn't register it fully. > Yes, I will add that to the new Jenkis setup asap. (some of the next days). > Thx, > > Sven > > PS: Another questions is: is 1.3 'released', if so where are the links on the website, where is the 'noise' ? We should fix the last open reports, create a new one-click and do the releaseâ¦ http://code.google.com/p/pharo/issues/list?can=2&q=milestone=1.3 -- Marcus Denker Do it together: http://taskfor.us

1 0

July 26, 2011

14057 ----- - Issue 4513: ScrollBar and MorphicModel needs cleaning. Stef and Igor killing tha big monster - Part two. :) http://code.google.com/p/pharo/issues/detail?id=4513 Gary it would be good that you check because we have two ways to to a perform: accessor and setValueSelector while only one would be enough. Stef and Igor

3 8