behdad
commented
5 years ago The reason HarfBuzz does it that way is because we found that closest to how Uniscribe does it.
Open NorbertLindenberg opened 5 years ago
behdad
commented
5 years ago The reason HarfBuzz does it that way is because we found that closest to how Uniscribe does it.
The USE documentation says, in the section “Default glyph pre-processing group”: “These features are applied together, one cluster at a time.” The documentation for a few other shaping engines also mention per-cluster application of features, and others seem to use it without ever mentioning it.
However, I can’t find any documentation specifying what per-cluster application really means, and there are differences between implementations in at least one respect: HarfBuzz requires the complete pattern of a rule to match characters within the cluster, while CoreText requires only the target of the rule to match characters within the cluster and allows the context to match characters before or after the cluster.
More precise documentation is needed.
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.