[PATCH 3/5] gas: partly restore how current_location() had worked

[Jan Beulich]

On 26.11.2024 03:44, Maciej W. Rozycki wrote:
> On Fri, 22 Nov 2024, Jan Beulich wrote:
> 
>> Commit 4a826962e760 changed its behavior without saying why, and without
>> putting in place any testcase demonstrating the required behavior.
> 
>  That comes from CVS days when change descriptions were not our policy (I 
> wish the policy was strictly followed nowadays too, sigh).
> 
>  Please refer these discussions:
> 
> <https://inbox.sourceware.org/binutils/alpine.DEB.1.10.1007240320070.29495@tp.orcam.me.uk/>
> <https://inbox.sourceware.org/binutils/alpine.DEB.1.10.1007241701380.29495@tp.orcam.me.uk/>
> <https://inbox.sourceware.org/binutils/alpine.DEB.1.10.1007241721120.29495@tp.orcam.me.uk/>
> <https://inbox.sourceware.org/binutils/alpine.DEB.1.10.1007242332460.29495@tp.orcam.me.uk/>
> <https://inbox.sourceware.org/binutils/alpine.DEB.1.10.1007260105250.29495@tp.orcam.me.uk/>
> 
> for the rationale behind this change and the other three in the series.  
> There is a test case for the whole series included in 4/4.  Arguably the 
> example given in 2.5/4 should have been made a proper test case as well.  
> Please feel free to use it now.

For the .eqv case current_location() keeps using dot_symbol, even though -
see the cover letter of the series - I'm unconvinced dot_symbol is being
maintained as it should be (or alternatively dot_frag / dot_value aren't).

To give some background: I first tried to actually get rid of dot_symbol
again, yet that didn't work out as easily as I had hoped.

Jan