[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Devel] FT_Set_Transform more info

From: David Turner
Subject: Re: [Devel] FT_Set_Transform more info
Date: Wed, 16 May 2001 23:04:19 +0200

Hi Mike,

Mike Owens a écrit :
> My problem seems to be in the cache manager. When I open the font library I 
> immediately cache the lower half of the character set. In one scenerio I was 
> setting the transform before characters were cached, hence, all characters 
> were rotated.
> In the other scenerio, characters are cached with transform set to 0 degress. 
> Then I set the transform 90 degrees. FTC_SBit_Cache_Lookup seems to realize 
> that lower case rotated characters are not in the cache but does realize the 
> upper case characters aren't. It doesn't seem to look at the transform which 
> may be the desired implementation.
> So now my question is, how do I design my program to be able to change the 
> current rotation of the current font face? Do I have to create separate font 
> faces for each rotation?
> I would appreciate some suggestions.

The cache sub-system simply isn't been designed to cache rotated glyphs,
mainly because it would have brought un-needed complexity to its design
(which isn't already so simple anyway).

It is possible to achieve what you want though, by cheating a bit. The
idea is simply to create, for each font face, one specific FTC_FaceID
per desired rotation/angle value (for example, 0 degrees and 90 degrees)

each FTC_FaceID will be resolved to a different FT_Face by your
face requester callback, where you'll appropriately call FT_Set_Transform
on the face object that is returned.

this should work transparently, with any kind of cache (though I suppose
you're using the small bitmaps cache). The only drawback is that you may
end up opening the same font file twice; however, this is not really a
problem since the cache sub-system already limits the number of opened

Hope this helps,


- David

reply via email to

[Prev in Thread] Current Thread [Next in Thread]