Re: mfc42: Stub implementation with the message where to get it from

[Alexandre Julliard]

Vitaliy Margolen  writes:
> This is more of the testing waters type of patch. If this approach is 
> fine then we can do the same for all other dlls that are shipped with
> Windows but not really a part of an O/S.

I don't think you can put up a message box on attach, that could
potentially become very annoying since it will trigger also for version
checks etc.  That sort of thing should be done at the loader level.

Recommending fixes (was: mfc42: Stub implementation with the message where to get it from)

[Paul Chitescu]

On Wednesday 09 September 2009 15:25:56 Alexandre Julliard wrote:
> Vitaliy Margolen  writes:
> > This is more of the testing waters type of patch. If this approach is
> > fine then we can do the same for all other dlls that are shipped with
> > Windows but not really a part of an O/S.
>
> I don't think you can put up a message box on attach, that could
> potentially become very annoying since it will trigger also for version
> checks etc.  That sort of thing should be done at the loader level.

What about a generic mechanism that would trigger a separate program to 
recommend - and possibly install - missing components? That program doesn't 
even need to be a win32 program.

My current idea is to somehow catch the fixme and err messages and pipe a copy 
of them to a filter that will identify "interesting" strings and build a list 
of fix candidates. Then it can start something like 
winetricks --prompt="Recommended installs to fix your problem" mfc42 vb6run

This way we may detect other problems that may have a known solution like 
missing DLLs, unimplemented OLE interfaces, assemblies with wrong version.