On 10/20/2013 3:45 PM, Philippe Verdy wrote:
2013/10/20 Asmus Freytag <[email protected] <mailto:[email protected]>>Incidentally, the dotted circle shown in the Unicode Code charts is *not* 25CC, and if I were to implement a "show dotted circle" feature in a program I would not use 25CC for this - that character has a standard glyph of rather unsuitable metrics for the purpose, never mind that many people have co-opted it.In fact the OpenType specs suggests assigning a glyph for the dotted circle so that renderers can use a glyph with the correct metrics for combining marks that are mapped in the font. The font may also assign distinct glyphs for some pairs with that base character, if the font supports multiple scripts.
If that is so, the it is an unfortunate problem with the OpenType specification.
It is not in agreement with the way U+25CC was encoded.
Many OpenType fonts are built like that so that the combining marks in isolation will not show on top of an unknown glyph with incorrect metrics, or worse over a zero-width space where they will collide with everything else on both sides.
Which is a limitation of the technology not based on conformance requirements by the Unicode Standard.
OpenType fonts however do not determine themselves is sequence is ill-formed : it is the renderer that parse these contexts and which then infer when to insert the base placeholder (it gets the glyph to insert by looking in the same font for the U+25CC mapping, otherwise it will use random or default or builtin font to get the dotted circle glyph, but it won't be able to correctly position that glyph and the combining mark from the font).
