[PATCH] malloc: Use compat_symbol_reference in libmcheck [BZ #22050]

[Florian Weimer]

On 08/31/2017 06:08 PM, Carlos O'Donell wrote:
> On 08/31/2017 11:04 AM, Joseph Myers wrote:
>> On Thu, 31 Aug 2017, Florian Weimer wrote:
>>
>>> On 08/31/2017 05:42 PM, Zack Weinberg wrote:
>>>> On Thu, Aug 31, 2017 at 11:37 AM, Carlos O'Donell <carlos@redhat.com> wrote:
>>>>> I would like to see a new macro that does what it says, rather than use the
>>>>> existing macro in the wrong way. Even if the new macro is just a copy.
>>>>>
>>>>> This looks like a real problem for glibc, particularly if we need to continue
>>>>> to use, at least internally, certain old versions of symbols. So having a
>>>>> new macro for this is fine.
>>>>
>>>> I see immediate uses for this macro in the test suite, verifying that
>>>> compat symbols continue to work correctly...  (particularly thinking
>>>> of the messy and totally untested old-FILE support).
>>>
>>> That's the exact purpose of compat_symbol_reference.  I think Carlos is
>>> objecting to its use for a *definition*.
>>
>> Well, I used it for the definitions of matherr and _LIB_VERSION in my
>> tests of those compat symbols, because it does exactly what's expected:
>> makes the definitions in the tests refer to the same entity as the compat
>> symbols in the shared libraries, rather than being completely independent
>> entities as they would by default.
>   
> While it does what's expected, the macro API wasn't designed with that in
> mind was it?

I would have used it in tst-mallocstate for __malloc_initialize_hook if 
I had understood what was going on.  I think the usages are so similar 
that we do not need a new macro.

Thanks,
Florian