Folks,

The problem Mark describes is certainly a bug/gotcha, but hardly a 
new one.  It has been present in all Smalltalk/Squeak browsers in our 
evolutionary line since the beginning of time (i.e. 1976) -- because 
so much of browsers' list-related logic is index-based rather than 
object-based.

The "smart updating" facilities added recently to Squeak reduce the 
incidence of such inter-browser conflicts by 99 percent, but do not 
(yet) catch the one that Mark mentions.

I can see that the "smart updating" has produced a new climate of 
positive expectation in which the remaining 
multiple-browser-coordination hazards, though they have been with us 
forever, seem new and unbearable!

That's progress, y'see.

   -- Scott

At 6:58 AM -0500 1/14/00, Mark A. Schwenk wrote:
>I really appreciate the attempts at browser synchronization, but 
>noticed this bug or gotcha in Squeak 2.7 update 1782.
>
>1. Open two browser windows on the same class category.
>2. Select a class and method in the first window.
>3. Change something in the method text.  Go to the second window and 
>add a class whose name preceeds the name of the class in the first 
>window.
>4. You notice a new red border in the first window.
>5. Activate the first window again and save the method. You are warned:
>Caution! This method has been changed elsewhere since you started 
>editing it here. Accept anyway?
>6. You are confused by the message and think the system must be 
>confused and choose yes. It will actually be saved in the class 
>preceding the class you begin editing in the list of classes. Ouch!
>
>If it isn't easy to fix the problem, perhaps we should reword the 
>warning message to something more generic like:
>
>Caution! This browser's contents are no longer valid due to changes 
>made in another browser. It is recommended that you open a new 
>browser, copy your changes to it, and save them in the new browser. 
>Accept anyway?
>
>-Mark Schwenk
> WellThot Inc.