[PATCH] ELF: Properly copy and strip Solaris binaries

[Rainer Orth]

Hi Alan,

>> Just a few comments and questions on the patch:
>> 
>> * The ChangeLog should refer to PR binutils/33718 so the commit shows up
>>   in there
>
> It is sufficient to put "PR 33718" in the subject.  Note that you
> need the space.

I see.  GCC is different here: the convention is PR<id> without a space.
However, the ChangeLog handling of the git commit hook checks both
syntax and completeness of the mandatory ChangeLog entries in the commit
message and extracts the PR references from that.

>> * I wonder why neither the PR nor the patch refer to the extensive
>>   discussion of the issue in PR binutils/33684 which prompted this PR
>>   and patch.
>
> Added to the patch.

Fine, thanks.

>> * Neither the patch names nor the filenames carry any meaning.  I think
>>   they should be way more descriptive instead.
>> 
>> * Besides, I wonder about dropping arbitrary objects into the testsuite.
>>   I suspect it would be way better to rely on native testing for issues
>>   like this, so we could both extend testing to all of Solaris/SPARC and
>>   x86, 32 and 64-bit.  Besides, it would be possible to run elfdump >
>>   /dev/null on the objects which is an excellent first sanity check for
>>   all kinds of objects on Solaris.
>> 
>>   After all, we're just one PR away from clean Solaris testresults as a
>>   prerequisite for setting up such bots :-)
>
> OK, I'll remove the testcase and just commit the objcopy.c part.

Thanks.  Ali already provided the native testcases in PR 33684.  I'll
just take it from there.

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University