On Wed, Sep 23, 2020 at 12:37:52PM +0200, Andreas Schwab wrote:
> On Sep 23 2020, Jakub Jelinek via Gcc wrote:
>
> > Even that 8.4.1 20200918 is redundant, r8-10517 uniquely and shortly
> > identifies both the branch and commit.
>
> But it requires a repository to identify. Often, redundant information
> is useful.
To identify what? What one usually does with gcc-testresults is look where
something regressed, for that because the short ids are monotonically
increasing, it is easy to find what is before and what is after.
And once the regression is narrowed to a particular commit (or some range of
them if not everything is covered), one needs the repository anyway,
version and date doesn't tell anything interesting, what one cares about is
what actually changed (and why).
Jakub
