[SMALL BUG]TranscriptStream in MVC?

Dan Ingalls Dan.Ingalls at disney.com
Fri Sep 15 18:59:02 UTC 2000


"G–ran" Hultgren <gohu at rocketmail.com> wrote...
>>> transcript := TranscriptStream new.
>>> transcript openLabel: 'Test'.
>>> transcript show: 'Why does not this text show up?';cr.
>>> 
>>> This opens the new Transcript but no text is shown.
>>> 
>>> transcript show: 'But this does!!!';cr.
>>> 
>...
>I replied:
>> You were only lucky if this worked in morphic.  I executed
>> Transcript cr; print: 3
>> with a transcript open, and saw nothing until I did
>> Transcript endEntry.

Henrik Gedenryd <Henrik.Gedenryd at lucs.lu.se> added...
>Dan: your example uses print: which merely puts text in the stream, without
>forcing it to the screen, whereas Göran's example uses show: which also
>forces the screen to update. (--literally--in Morphic it sends a message to
>update the whole screen, a rather expensive operation, but perhaps necessary
>to ensure consistent behavior in certain debugging situations.) So the
>behavior in your example is not a bug. Göran's problem seems unrelated. (And
>the only way to make a final #cr or #tab have an effect is by sending
>endEntry, or doing another show:)

I carelessly missed the fact that *both* his examples used show: 
I intentionally used print as an example of something that doesn't properly
send endEntry, since I (mistakenly) thought that was the real topic.

Anyway:
1) if anyone wants to fix G–ran's problem with MVC, that would be great ;-).

2)  I don't think the performance penalty of endEntry is that bad.
The morphic message to "update the whole screen" is named that way,
but it only really redraws the damaged regions.

>Perhaps changing to a stepping updating mechanism would be the way to go in Morphic?

My thoughts exactly.  Try it out.

	- Dan






More information about the Squeak-dev mailing list