Subject: Re: Win32 performance results
From: WJCarpenter (bill-abisource@carpenter.ORG)
Date: Wed Mar 07 2001 - 12:54:10 CST
>> Can we do something about all the calls to GR_Graphics::remapGlyph?
>> It calls XAP_Prefs::getCurrentScheme and XAP_App::getPrefs twice
>> for each invocation.
paul> These both sound like places where some intelligent caching
paul> could help. Of course, all of the usual cache maintenance
paul> caveats apply.
>> Can we do something about the gazillion calls to remapGlyph and
>> measureUnRemappedChar?
A while back on the list, there was a discussion about a text run
having a shadow copy of the text to be used just for rendering (the
discussion arose, AFAIR, from a discussion of how the various
rendering engines use different character sets and encodings other
than our 16 bit stuff). In a nutshell, it would have these kinds of
characteristics:
1. Produced on demand at first render time if the rendered text was
in some way different from the actual text (for example, remapped
glyphs are such cases, whereas smart quote substitutions are not).
2. If the text in the run were altered in any way, the shadow copy
would be discarded. Then, see step #1.
I'm omitting lots of details, of course, that would need fixing up by
an implementer, but I think that would help in the particular cases
mentioned above. I had been re-pondering this strategy while
considering enhancing remapGlyphs to be able to do font changes. That
would certainly make it more expensive. Once implemented, for
whatever motivation, I think these would find many uses.
-- 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 : Wed Mar 07 2001 - 13:52:46 CST