Abdelrazak Younes wrote:
Jean-Marc Lasgouttes wrote:
"Abdelrazak" == Abdelrazak Younes
<[EMAIL PROTECTED]> writes:
Do we really want to use a vector to cache unicode data?
Abdelrazak> No, a vector to cache width (int) data. This is to replace
Abdelrazak> the map font width, nothing to do with unicode.
How large will this vector be?
Abdelrazak> As large as the number of character width and font as the
Abdelrazak> document use.
That was my question: isn't the font supposed to have 2^16 (2^32?)
characters?
Ah, sorry about the misunderstanding. IIRC correctly there was a trick
in Andre solution to avoid that big size. Need to search the archive.
There was no such trick. Andre proposed to limit the cache to the first
256 chars. I pushed that to 65536 (2^16) which covers all of ucs2. Even
if we use 100 different fonts and sizes in a document this add less than
7 megs of memory.
Abdel.
