Issues with ro_Latn checks

[moyogo]

ro_Latn.yaml currently defines two shaping_differs checks which check default Ş and Ş in Romanian differ or default ş and ş in Romanian differ.

There are a few issues here:

1. If Şş should look different in Romanian, so should Ţţ.
2. There are use cases where some fonts have a cedilla that looks like a comma and may use the same component for ȘșȚț and ŞşŢţ, the check won’ be helpful.
3. ȘșȚț were added in Unicode 3.0 in 1999 for Romanian, there are use cases where ŞşŢţ should look like ȘșȚț (pre Unicode 3.0 or ISO/CEI 8859-2 data which was used well beyond 1999) and there are uses cases where ŞşŢţ should not look like ȘșȚț (post Unicode 3.0 or ISO/CEI 8859-16 data used from 1999). For example, one may find it useful to be able to tell if a text needs to have ŞşŢţ replaced by ȘșȚț, or when Gagauz or Turkish names with ŞşŢţ are shown in Romanian/Moldavian text.

Although ȘșȚț have been available for 25 years, considering this is still a common practice and both designers and users may or may not expect this, this should be a warn instead of a fail.

ShapingDiffersCheck would need to be able to warn as well as okay or fail.