Subject: Re: RemapGlyph()
From: WJCarpenter (bill-abisource@carpenter.ORG)
Date: Thu Jun 21 2001 - 13:23:37 CDT
wjc> So, in summary, I think the root cause of bug 1518 is in whatever
wjc> thing created the sample document attached to the bug, or
wjc> possibly the importer code is not doing the right thing.
andrew> They're in the U+0300 to U+0345 range. Yep the combining
andrew> diacriticals. Abi does have problems. Static rendering is
Let me clarify my previous statements. When I said "remapGlyphs is
not the cause of the problem in bug 1518", what I meant was
"remapGlyphs is exactly the cause of bug 1518". It's a subtle
difference. :-) If you turn off remapGlyphs (set preference item
"RemapGlyphsMasterSwitch=0"), the Vietnamese text in the bug's sample
document is rendered correctly (at least on MS W2k, and at least as
far as my guesswork about Vietnamese combining diacriticals will
carry me).
As Andrew Dunbar points out, it is wrong to treat "missing glyphs" and
"zero-width glyphs" as the same thing. The trouble is, I don't think
you can tell the difference in any reliable way on some of the
platforms (maybe I'm wrong about that, though). We have so many
different font things going on, so much swearing about fonts, etc,
that I suggest the best course is to have a new method with a catchy
name like "glyphExistsForCodepoint(...)" and have remapGlpyhs depend
on that instead of depending on the font metrics directly. For places
where we can't do better then guessing from the metrics, those places
will continue to be lame while the other places can correctly answer
the question.
That means a new method for each different sort of font technology,
but the problem for which remapGlpyhs was invented still exists. So,
better to fix up this bit of business with fonts than to get rid of
remapGlyphs.
I won't have any chance to look at this in any predictable timeframe,
so any eager beaver who thinks they can handle it should dive in with
at least the goal of getting past bug 1518, even if all the other
points about zero-width glyphs that Andrew mentioned can't be handled
until later.
-- bill@carpenter.ORG (WJCarpenter) PGP 0x91865119 38 95 1B 69 C9 C6 3D 25 73 46 32 04 69 D6 ED F3
This archive was generated by hypermail 2b25 : Thu Jun 21 2001 - 13:53:14 CDT