This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] [BZ #16512] Strip @@GLIBC_.* and @@GCC_.* from symbol names


On Wed, Jan 29, 2014 at 2:09 PM, Roland McGrath <roland@hack.frob.com> wrote:
>> On Wed, Jan 29, 2014 at 1:23 PM, Roland McGrath <roland@hack.frob.com> wrote:
>> >> "readelf -r" may dislay symbol version in dynamic relocation:
>> >
>> > When does it and when doesn't it?
>>
>> If you use binutils with the fix for
>>
>> https://www.sourceware.org/bugzilla/show_bug.cgi?id=16496
>>
>> like Linux binutils 2.24.51.0.3, you will get symbol version
>> in dynamic relocation.
>
> I see.  So the answer is that it doesn't ever do it, and you've proposed
> making it always do it, but that has not been approved by binutils
> maintainers.  (Breaking existing scripts this way might be a good reason
> the change gets rejected.)

The patch wasn't rejected.  There are just no comments.  Dynamic
relocation without symbol version doesn't provide complete information
to tell if there is a problem.

>> > Perhaps it would be better to have the exact symbol version be part
>> > of each reference list.  But perhaps it would be worse.  I have not
>> > given it a lot of thought.
>>
>> It won't work with the binuitls which doesn't have the PR 16496
>> fix.
>
> There are multiple ways to get the information.  That will not be a barrier.

Sure, apply my fix is one of them.

> We just have to decide whether we think it's the right thing to do or not.
>


-- 
H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]