[binutils-gdb] gas: partly restore how current_location() had worked

[Maciej W. Rozycki]

On Sun, 12 Jan 2025, Maciej W. Rozycki wrote:

> > > Well, maybe. I'm sorry if you feel offended. I don't, however, think there
> > > ever was a policy to explicitly forbid describing changes beyond their
> > > mechanics.
> > 
> > I think, in that vein, it would have been worth including a link to the
> > elucidative ML discussion. No policy against that? ;)
> 
>  In the CVS days the policy was to put ChangeLog entries and nothing else 
> in the commit log.  I don't think we have a policy nowadays though to put 
> inaccurate or plain incorrect information in git change descriptions.

 NB coming from the Linux kernel world I was quite unhappy with the old 
policy and also an advocate for and an early adopter of meaningful change 
descriptions (which can be easily determined by examining git histories of 
binutils-gdb and gcc repos), so I feel double-insulted with persistent, 
not taken back in the light of my explanation given, and now recorded for 
posterity in the git history an accusation of using an inadequate change 
description while indeed it was exactly the requirement at the time.  And 
also the unfounded claim as to the lack of a test case despite that there 
has been one included at the time and I did point at it now.

  Maciej