Re: [Q] documenting font construction

It is just a matter of behavior - documentation match. After all, shell commands like one described in "Font Instancing" in the patch: xlsfonts -fn "$instantiator" | head -$MAX_FONT_COUNT | grep "$charset_registry" | head -1 helped me tremendously to diagnose and understand the issue. To explain why `grep' argument is as stated, one needs a reservation about how charset registry is matched against in X. When XLFD format is always used for font names, the visible difference in behavior is really unlikely. But font names do not have to be in that format, and (contrary to what "Face Resources" currently says) not on all systems all X fonts can be referred to via XLFD-style names. In particular, on my sco 5.0.2 `8x14pc437', `8x14pc850', `8x14pc932' have no XLFD aliases. Do not using such a fonts currently for Mule, but can easily imagine other users doing so. So description of matching behavior on font name in any format is required. What happens if non-XLFD font names match charset registry currently is unclear, but made clear with the patch, and must remain clear whatever `x_find_charset_font' re-implementation will be. The total exlusion of non-XLFD fonts from matching is undesirable, so any re-implementation must provide a trapdoor to use non-XLFD font names after all XLFD ones are exhausted. Since we already refer to XLFD, avoiding X-dependent language in documentation is hard to impossible anyway.