Re: [3/6] widl: Avoid generating negative offsets for conformant arrays of strings. Take 2.

[Alexandre Julliard]

Dmitry Timoshkov  writes:
> This version of the patch limits the changes to conformant arrays of
> strings in order to avoid unwanted changes to other pointer types.

Didn't you say that midl is doing this for all arrays? Why doesn't this
work in our case?