Apache OpenOffice (AOO) Bugzilla – Issue 7004
Re-encode known symbol/dingbat/graphical/etc. fonts to use correct ISO-10646/Unicode instead of using 0xF020 to 0xF0FF.
Last modified: 2003-01-27 17:06:12 UTC
Some known symbol/dingbat/graphical/etc. Fonts (eg: "Zapf Dingbats", "StarBats", "StarMath", etc.) which has characters in ISO-10646/Unicode, should be re-encoded to use correct ISO-10646/Unicode characters rather than the generic 0xf020 to 0xf0ff range which are used by (generic non-textual) graphical fonts. Although I have noticed "fontcvt.cxx" at "http://gsl.openoffice.org/source/browse/gsl/vcl/source/gdi/", when I open up the Special Character box in OpenOffice.org 1.0.1, I see fonts such as "Zapf Dingbats", "StarBats", "StarMath" still use the 0xF020 to 0xF0FF range. But I found that Ghostscript's "Standard Symbols L" is re-encoded(?) correctly.
sorry, I can't see the issue. What encoding would you expect ? What has fontcvt.cxx to do with it ?
I have not said that there is any problem with that piece of code that I have located, but it seems to be related to this feature request. -- A database of Symbol Fonts with their ISO-10646/Unicode mappings. I meant to include a feature to re-encode the known graphical, dingbat, Greek or whatever (SYMBOL) fonts which contain characters in the ISO-10646/Unicode to use the ISO-10646/Unicode encodings, rather than using the 0xF020 to 0xF0FF encoding which is used by all SYMBOL fonts regardless if they have known mappings or not. If the SYMBOL font's name is known in the mapping database: -> Remap the font according to ISO-10646/Unicode. -> Otherwise use the default 0xF020 to 0xF0FF.
If the supplied encoding of the font is wrong, then I recommend to fix the font. Providing a databased reencoding will lock out all customer that try to substitute a reencoded font with an other one. Even supplying the same font under a different name will break the document. Furtheron documents created with OOo will not be readable with other applications (e.g. Mozilla) even when using the same font.
I think that you may have missed the point: I have many non ISO-10646/Unicode compliant Graphical/Dingbat fonts, & I don't have any program to reprogram the character maps inside font files. The ISO-10646 feature makes it easier for me to change fonts in documents without the problems that is associated with legacy 8-bit (pre-Unicode) word-processors. If an user does not desire that Graphical/Dingbat fonts to be remapped (for some compatibility reason) according to the built-in databases, add a GUI option to disable this behaviour. "Disable Unicode remapping of Symbol Fonts" Also, you can leave the symbols at 0xF020 to 0xF0FF, but mirror these glyphs to their correct corresponding ISO-10646/Unicode positions (to be compatible with both ISO-10646/Unicode and older software). Also, I would like to have ability to do user defined character map redefinition on any arbitrary font, even it overrides the built-in databases. I have some non compliant forgein alphabetic fonts which I would like to re-encode. I have tons of legacy (non-Unicode aware[?], & also pre-Euro) TrueType fonts with fi/fl ligatures that are encoded in the WRONG positions at 0xF001 & 0xF02 (in the Private Use) rather than the standard 0xFB01 & 0xFB02. Also, if a font has glyphs at 0xF001 & 0xF002 (badly encoded fi/fl ligatures) but no glyphs at 0xFB01 & 0xFB02, mirror (or remap) the 0xF001 & 0xF002 to the correct 0xFB01 & 0xFB02 so the fi/fl ligatures can be used normally. A GUI option can be added for the above "Mirror badly encoded fi/fl ligatures to correct Unicode codepoints".
closed