[Qgis-psc] Attachment widget broken in the attribute table

[Sandro Santilli]

On Tue, Aug 06, 2019 at 11:55:28PM +0200, Jürgen E. Fischer wrote:

> I'm always puzzled with the obsession on regressions.  To me it doesn't matter
> much whether a bug is a regression or not - it's only the impact of a bug that
> counts.  A regression can have big impact - but if it doesn't it should not be
> a priority over bugs with bigger impact.

I think the impact of a regression is by definition much higher than
the impact of a non-regression bug. Known bugs users can work around,
regressions are really bumps users are faced with. Unexpected things
that stop working on upgrade. They make people afraid of upgrading.

That said, people (like me, which is a big surprise, being a
programmer...) that prefer to avoid such bumps can stay to LTR
releases, and in LTR releases regressions would seriously be of
detriment of the reputation of the project. Regressions in major
releases can be tolerated more (as long as they get fixed before
the new version becomes LTR...)

--strk;