Re: Reverting unit tests?

On Fri, Nov 25, 2016 at 10:17:46AM +0100, Jan Holesovsky wrote:
> Hi Khaled, Miklos,
>
> Khaled Hosny píše v Čt 17. 11. 2016 v 15:22 +0200:
>
> > It seems to not cover all the symbols in Symbol (what a great font
> > name), or the original bug fixed by the commit in
> > https://gerrit.libreoffice.org/#/c/30932/ wouldn’t happen in the first
> > place.
>
> I've just seen this - I am not sure it's a good idea to remove a unit
> test...
>
> In the commit message there is "The bug doc is broken in master with or
> without this commit, but reverting this fixes tdf#103944." - I'd prefer
> keeping the test in, and just #if 0 it until we have a fix; ideally with
> bug number in a comment for a bugreport that is marked as regression
> etc. so that it is on the radar / ESC stats?

The thing that the unit test was checking (that RTL_TEXTENCODING_DONTKNOW
is used) caused another bug and IMO was the wrong fix, so keeping the
unit test did not make much sense. Now thinking about it, may be I
should have kept the test and made it check for RTL_TEXTENCODING_SYMBOL
instead, right?

The original issue is now fixed in 37a14c5e359a554195fbcbe2805a284ccf4cb677,
but no unit test there since the fix is now in the font and don’t know
how to test that.

Regards,
Khaled