This is the mail archive of the
binutils@sourceware.org
mailing list for the binutils project.
Re: Three more days left before the 2.30 release...
On Fri, Jan 26, 2018 at 10:50 AM, Mark Wielaard <mark@klomp.org> wrote:
> Hi Nick,
>
> On Fri, 2018-01-26 at 17:51 +0000, Nick Clifton wrote:
>> Right, well given how close we are to the deadline for the 2.30
>> release, and the fact that it does not look like this issue is
>> going to be resolved quickly, I am going to have to make an
>> executive decision.
>>
>> I am going to stay with the sources as they are for now.
>
> I appreciate that you'll have to make a decision to fix things or keep
> the existing code because it is too close to release time. It is
> ultimately your decision. But...
>
>> Sorry
>> Mark, but I without any concrete proof of broken tools, I am
>> going to trust that H.J.'s change will work out in the long run.
>
> I think this is not really fair. It should be on the person proposing a
> new, incompatible, format to argue for why it is essential, not on the
> people having to fix the tools to support the new format, to proof such
> a breaking change is necessary.
>
> I have been raising this issue since at least November last year on
> various gnu-abi/gabi lists. I believe the consensus has always been
> that changing the GNU ELF notes format after 20 years is just asking
> for trouble. That is also what people said reviewing H.J.'s binutils
> and glibc patches. It complicates anything dealing with ELF notes since
> there are now suddenly multiple different formats to support.
>
> Also it is fairly easy to show "concrete proof" of broken tools. H.J.
> had to patch binutils itself, glibc and the linux kernel to just parse
> the alternative format. I have fixed some things in elfutils to at
>
This is not entirely true. The binutils and glibc can parse the standalone
property note just fine since they don't need note alignment to get the
next note. What I did was to make up a new note with 8 byte alignment
and placed it before the property note by hand. That is how I found the
problem.
--
H.J.