On Thu, Oct 28, 2010 at 11:27 PM, Adrian Lienhard <[email protected]> wrote:
> Ok, this is not related to closures. Polymorph uses blocks, e.g., for the > addUpAction of a MultistateButtonMorph (in this case the block is created > in UIThemeWatery>>#createCloseBoxFor:). And blocks are not serializable with > reference streams. Therefore, you cannot write out a window that has > buttons. The original buttons in Morphic used action selectors instead of > blocks. It shouldn't be hard to refactor Polymorph buttons to take the > receiver, selector, and possible arguments instead of a block but I wonder > whether its worth the trouble? > But in principle there's no reason why BlockCLosures and their associated home contexts and methods can't be stored on a DataStream. Is anyone brave enough to try and eliminate the restriction? best Eliot > Cheers, > Adrian > > > On Oct 15, 2010, at 02:22 , Sheridan Mahoney wrote: > > > > > Hi all, > > The method that generated the error in issue 3090 is below: > > > > > > MethodContext(ContextPart)>>storeDataOn: aDataStream > > > > "Contexts are not allowed go to out in DataStreams. They must be > included inside an ImageSegment." > > > > aDataStream insideASegment ifTrue: [^ super storeDataOn: > aDataStream]. > > > > self error: 'This Context was not included in the ImageSegment'. > > "or perhaps ImageSegments were not used at all" > > ^ nil > > > > > > I've been trying to make sense of the stack, and it appears that writing > a morph > > to a file is hooking an object which cannot be written to a DataStream. > > > > Does anyone know whether the proper behavior is to: > > - have the MethodContext go out without being inside an ImageSegment, or > > - fix the morph-writing so that it uses ImageSegments??? Mariano would > love > > this, right? :) > > > > Or perhaps there's a better way to understand the stack. > > > > Many thanks, > > and Cheers, > > > > Sheri/dan > > > > > > *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-**-*-*-*-*-* > > Listening's Auspicious Fortune is a heart > > Rejuvenated with Profound Instructions. > > > > Reflection's Auspicious Fortune is the collapse > > Of delusion's false cavern. > > > > Kabje Kalu Rinpoche > > > > "With a pure motivation that does not wane and > > with great courage that does not despair, may > > each of us endeavour to make our lives meaningful " > > > > H.H. the 17th Karmapa > > Orgyen Trinley Dorje > > > > > > > > > > On Oct 11, 2010, at 7:16 AM, Sheridan Mahoney wrote: > > > >> I've just re-entered the Pharo world after a long(ish) absence, and > wanted to report a > >> new bug submission; here's the breakdown: > >> > >> Pharo image: dev > >> Pharo core version: Pharo-1.1.1-- Latest update: #11414 > >> Virtual machine used: 4.0.3.2202 Unix VM > >> OS: Ubuntu Lucid Lynx 10.04 > >> Class browser used if applicable: SystemBrowser default > >> > >> Steps to reproduce: > >> 1. start pharo > >> 2. open a Shout workspace > >> 3. select the just-opened workspace > >> 4. Ctrl-right button to get the meta-click window > >> 5. select 'save morph in file' > >> 6. hit 'ok' to proffered name for file > >> 7. a message box comes up that reads 'please wait while objects are > >> counted'. > >> 8. shortly after that an error is generated: > >> 'Error: this context was not included in the ImageSegment' > >> > >> > >> I'm planning on doing more research, but wanted to get the issue > reported. > >> > >> Cheers, > >> Sheridan > >> > >> PS - so far I really like 1.1.1 !!! Good work, gang! > >> > >> _______________________________________________ > >> Pharo-project mailing list > >> [email protected] > >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > > > > > > > _______________________________________________ > > Pharo-project mailing list > > [email protected] > > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > >
