Re: [lmi] Preserving configurable settings after GUI test

[Greg Chicares]

On 2018-07-30 15:19, Vadim Zeitlin wrote:
> On Mon, 30 Jul 2018 14:36:34 +0000 Greg Chicares <address@hidden> wrote:
[...]
> GC> As for PR 88, I've committed something simpler--sorry, I should have
> GC> realized that what I originally asked for was a lot bigger than what
> GC> we really need.
> 
>  Well, I did have my doubts about it, but finally decided it was better to
> be thorough for the reasons I explained in the commit message. Maybe I was
> too much so, but I still think your changes go too far in the other
> direction as they don't even restore the file in the case of test failure.

True. However, end users don't even know this code exists, so it doesn't
have to be really robust. Only three people in the world ever run these
GUI tests, and I don't think I've ever seen this particular test fail.

> I.e. if nothing else, I'd at least use RAII/scope guard/explicit try/catch
> to ensure that the 2 settings we want to preserve here are indeed preserved
> in this case.

I don't think it's worth any more effort at all. If it ever fails, we can
just re-select the settings we want, or restore them from a backup copy.