we really need help

[Andreas Raab]

Stef,

Just like with change sets (which would suffer from the same ordering 
problem of #insetBy: vs. isRectangle) you can't just throw all of these 
changes in the same pot and hope it'll work. You need to be a bit more 
orderly than that.

In this case, the solution is to make package version(s) which *only* 
includes the method #isRectangle; then post a configuration that ensures 
this package version is loaded; then commit the other fixes. That's 
basically the same way you'd use change sets - first make one that 
includes #isRectangle, then post the remaining changes.

There is simply no way by which MC could decide how to load these 
methods in the proper order. Sometimes you do need to give a helping 
hand when migrating a running system and that's one of the places.

Cheers,
   - Andreas

stéphane ducasse wrote:
> Hi all
> 
> We ***really*** need help. I start to be really exhausted. Here is  the 
> problem:
> 
> Marcus introduced a fix to speed up morphic isRectangle instead of  
> isKindOf: in insetBy:
> and collected some other fixes.
> 
> Now when we take a 6693 image
>     - merge MC275
>     - load scriptLoader-md.6.mcz
>     - execute script 6
> 
> you end up after merging to an infinite loop because SmallInteger  does 
> not know isRectangle
> (even thought isRectangle is defined on Object). This sounds like MC  
> loaded first the method
> insetBy: and that the progress bar needs it and Object>>isRectangle  is 
> not loaded yet.
> So what can we do?
> 
> I tried to tweak a bit the script by loading first kernel-md.42.mcz
> take a 6693 image
>     - merge MC275
>     - scriptloader-sd.8.mcz
>     script 6
>     - but now I get an error because Debugger does not know stepAt:in:
> 
> So I do not know what to do except ordering things by hand.
> I lost really a lot of time and be completely unproductive.
> So does anybody have any suggestion?
> 
> Stef
> 
>