[3/3] Adjust testcases for readelf printing 'foo@@'

[Michael Matz]

Hello,

On Thu, 20 Nov 2025, Jan Beulich wrote:

> On 20.11.2025 16:30, Michael Matz wrote:
> > those are the public base-version symbols.
> > ---
> >  ld/testsuite/ld-elf/pr23161b.rd           | 6 +++---
> >  ld/testsuite/ld-elf/pr25617-1a-sec-hdr.rd | 4 ++--
> >  ld/testsuite/ld-elf/pr28348.rd            | 2 +-
> >  ld/testsuite/ld-ifunc/pr23169a.rd         | 2 +-
> >  ld/testsuite/ld-ifunc/pr23169b.rd         | 2 +-
> >  ld/testsuite/ld-ifunc/pr23169c.rd         | 2 +-
> >  ld/testsuite/ld-plugin/pr22983.1.d        | 2 +-
> >  ld/testsuite/ld-plugin/pr22983.2.d        | 2 +-
> >  ld/testsuite/ld-plugin/pr22983.3.d        | 2 +-
> >  ld/testsuite/ld-plugin/pr22983.4.d        | 2 +-
> >  10 files changed, 13 insertions(+), 13 deletions(-)
> 
> Doesn't this need folding into one of the earlier patches, to avoid 
> intermediate fallout?

If patch two is _not_ approved then this patch would need either folding 
into the first one or must be committed before the first one (it will work 
before).  If patch 2 is approved then [3] is not needed at all, and I'll 
commit patch 2 before patch 1 to avoid intermediate fallout.

> > --- a/ld/testsuite/ld-elf/pr23161b.rd
> > +++ b/ld/testsuite/ld-elf/pr23161b.rd
> > @@ -1,7 +1,7 @@
> >  #...
> > - +[0-9]+: +[a-f0-9]+ +0 +NOTYPE +GLOBAL +DEFAULT.* [0-9]+ +_?_(_bss_start|edata|end)
> > + +[0-9]+: +[a-f0-9]+ +0 +NOTYPE +GLOBAL +DEFAULT.* [0-9]+ +_?_(_bss_start|edata|end)@?@?
> 
> Is it really correct to expect zero, one, or two @-s here? Don't you 
> mean zero or two?

True.  Consider it amended by '@?@?' --> '(@@)?' .  But I really really 
hope patch two makes all of this unnecessary.


Ciao,
Michael.