Re: [PATCH 2/2] LD/testsuite: Expand STB_GNU_UNIQUE test coverage

["Maciej W. Rozycki" <macro@imgtec.com>]

Hi Nick,

> > So does anybody actually know why the current per-architecture approach 
> > was accepted in the review?
> 
> No.  My guess is that it was not thought through properly.  Possibly a
> developer wanted it for their architecture, but did not consider the impact
> on other architectures.  So their patch went in for their architecture and
> the feature proved to be useful.  Then users for other architectures (probably
> kernel and library builders) decided that they wanted it too, and so a
> piecemeal set of hacks were added to accommodate them.  Just a guess mind...

 Thank you for your confirmation, I suspected this might have been the 
case.

> >  If not then I'll post a proposal separately to gather all the bits 
> > scattered across target `*_add_symbol_hook' handlers and handle this 
> > feature in `elf_link_add_object_symbols'.
> 
> Please do.

 I've got the test suite part done already, so I'll update code and post 
the two changes together.  I think it'll be the best if the code update 
goes in first or both are a single change, so as not to cause regressions 
which might break bisection.

  Maciej