[squeak-dev] Improving the SystemNavigation browseMessageList... stuff

tim Rowledge tim at rowledge.org
Sat Oct 17 20:37:28 UTC 2020


I just had reason to take a look at the list of unimplemented messages in my working image. I was a bit surprised to see such a large number *not* related to my current work - in fact the latest trunk image claims *577* cases. Obviously a few are legitimate places where some test needs an unimplemented symbol to find but still, that's a lot of places where sometihng is going to go wrong.

It has to be said that VeryPickyMorph>>#justDroppedInto:event: is especially amusing in this regard. Maybe not so picky after all?

What struck me though was how unhelpful the browser ends up being in this case. What we get is a long list of methods where
a) there are duplicates - see MenuMorph>>#setTitleParametersFor: for an example. It apparently sends titleColor, titleBorderStyle, titleBorderColor and titleBorderWidth and appears 4 time in the browser. I'd claim that it would be far better to build some variety of Set of MethodReferences in MessageSet>>#initializeMessageList:

b) we lose the useful info about which message is unimplemented in the parsing in MessageSet>>#initializeMessageList:. Oddly, I'd swear I've seen something, somewhere, sometime, that could holds onto some more info to pass onto a browser. Anybody? And of course, the browser would need some extensions to actually display the unimplemented message. This might even be a place where Shout would actually appeal even to me.

The first question really has to be whether anyone has already tackled this stuff?

tim
--
tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
A computer's attention span is only as long as its extension cor.................




More information about the Squeak-dev mailing list