[Glass] Grrrr cannot migrate (class rename with subclasses and with a name of a deleted class)

[Dale Henrichs via Glass]

To be even more clear ... if "you" expect me to "fix all of the loading 
problems" in a vacuum then it will not happen ...

OTOH. If someone takes the time to create a specific set of test cases 
that fail in the latest GsDevKit/GLASS then I am willing to work on them 
...

The moving variables up issue will probably require a fairly major 
rewrite of the Monticello loader (without wading into the problem I 
can't guess as to the magnitude of "major rewrite") and my primary 
concern is that you guys are hitting other corner cases that I am not 
aware of so I want to make sure that I am aware of as many of the 
problem areas as possible.

A failing test case is the best way to communicate that to me ...

Dale

On 8/28/15 11:42 AM, Dale Henrichs wrote:
>
>
> On 8/28/15 11:30 AM, Mariano Martinez Peck wrote:
>>
>>
>> Can you confirm that turning OFF automigration solves the example of 
>> moving instVars up?
>>
>
> Sorry if I implied that automigration implied the specific case of 
> moving instVars up, but I am also under the impression (possibly 
> mistaken) that the problems that prompted Otto to run that script were 
> not exclusively due to moving class vars up ...
>
> In other words, I don't believe that turning off automigration solves 
> that particular problem ... I don't have an automatic answer to that 
> problem ... because of the limitations of the current Monticello 
> loader algorithms ...
>
> I do believe (possibly mistaken) that there are additional 
> refactorings that can cause problems ...
>
> Dale