[PATCH] Follow-up to r206457 -- fix static adjustments for some subtle virtual inheritance cases

[Timur Iskhodzhanov]

Well, the ball has started rolling already, so I finished splitting
the nonvirtual tests in r206782.

I'll defer updating the virtual inheritance tests – there are more of
them, they are more complex and the task is indeed boring.
Agreed that the new tests should use the new format.

2014-04-18 22:57 GMT+04:00 Reid Kleckner <rnk at google.com>:
> No worries, I wouldn't spend much time rewriting old tests.  I'd write new
> tests this way, though.
>
>
> On Fri, Apr 18, 2014 at 11:47 AM, Timur Iskhodzhanov <timurrrr at google.com>
> wrote:
>>
>> OK, I'll continue this boring task next week then :D
>>
>> 2014-04-18 22:44 GMT+04:00 Reid Kleckner <rnk at google.com>:
>> > On Fri, Apr 18, 2014 at 8:17 AM, Timur Iskhodzhanov
>> > <timurrrr at google.com>
>> > wrote:
>> >>
>> >> 2014-04-18 1:38 GMT+04:00 Reid Kleckner <rnk at google.com>:
>> >> > ================
>> >> > Comment at:
>> >> > test/CodeGenCXX/microsoft-abi-vtables-virtual-inheritance.cpp:459
>> >> > @@ +458,3 @@
>> >> > +namespace Test12 {
>> >> > +struct X : B, A { };
>> >> > +
>> >> > ----------------
>> >> > Going forward, I'd really prefer it if our inheritance hierarchy
>> >> > tests
>> >> > could all be standalone.  See the microsoft-abi-vbtables.cpp test
>> >> > cases,
>> >> > where the contents of any namespace can be cut-and-paste into a
>> >> > standalone
>> >> > .cpp file and compiled.
>> >> >
>> >> > In this test, I have to go lookup ::A and ::B in the right namespace
>> >> > to
>> >> > figure out what kind of record this is going to be.  It's worth a few
>> >> > lines
>> >> > of duplicated test code to avoid that.
>> >>
>> >> Like r206614?
>> >
>> >
>> > Yep, thanks!
>
>