[PATCH] fix #19444 - build failures with -O1 due to -Wmaybe-uninitialized

[H.J. Lu]

On Thu, Jan 14, 2016 at 1:00 PM, Joseph Myers <joseph@codesourcery.com> wrote:
> On Thu, 14 Jan 2016, H.J. Lu wrote:
>
>> > Suppose you apply all Wilco's patches related to string inlines (including
>> > those that remove many of them from bits/string2.h).  Are there any
>>
>> His patch can't be applied since it contains odd characters.
>
> Well, the patches should be reviewed.  Presumably he has clean copies of
> them.
>
>> > remaining references to _STRING_ARCH_unaligned in installed headers other
>> > than the definition?
>> >
>> > If so, what are they?  If not, it should be moved out of installed headers
>> > as I said (and bits/ is a namespace exclusively for installed headers).
>>
>> crypt/md5.c:#if !_STRING_ARCH_unaligned
>
> This is a list of references outside installed headers.  What I want is a
> list inside such headers, after all relevant patches have been applied.

-Os problem is reference of _STRING_ARCH_unaligned in
glibc, which isn't defined with -Os.  If it is allowed to use in
glibc, its definition should always be available for glibc build.

> For example: there are lots of references in the installed header
> bits/string2.h at present.  So we can't move the macro out of installed
> headers until those references go away.  Does
> <https://sourceware.org/ml/libc-alpha/2015-12/msg00386.html>, together
> with <https://sourceware.org/ml/libc-alpha/2015-10/msg00265.html>, move
> all those references to a .c file or not?

They look promising.  I will wait until they are settled down.


-- 
H.J.