Re: [Freetype] the joys of FT_Set_Char_Size()

[The Rasterman]

On Fri, 23 Jan 2004 12:07:40 +0000 address@hidden babbled:

> Hi,
> 
> > To share a FT_Face object between multiple instances of that font face at
> > differing sizes...
> This is possible.  You can have many FT_Size objects (representing a
> font at a specific size) for a single FT_Face.  See:
> http://www.freetype.org/freetype2/docs/reference/ft2-sizes_management.html

aha!! read.. read... yes! bingo! just what i need! i shall give this a burl! :)

> >judicious use of a face cache and FT_Set_Char_Size().
> If you are using a face cache, why not just use FTC_Manager_Lookup_Size()?

i'm not using ft2's cache system as at the time i wrote the code the
documentation on ft2's cache manager was sparse and mostly non-existent. it was
listed as "a work in progress and will change", so i played a bit with tyring to
get it up - it didn't work, so i just did my own using my existing support
primitives. i probably would need these anyway if i were to support back-ends
other than ft2, so i saw no major loss in doing it. ft2's cache system may be
better now though (ie documented and stable) but too late now :)

> This should take care of caching FT_Size objects for you.  It also
> activates the FT_Size for you, so you can just use the FT_Face as normal.
> This is how I handle kerning.

yup yup. cool- but the above solution is just what i need :) shouldn't take long
to squeeze into my existing api :)

> Kind regards,
> 
> Jon Foster
> --
> 
> Carsten Haitzler (The Rasterman) <address@hidden> wrote:
> 
> Ahh and so at long last it is that I would go through my code and fix a FIXME:
> that has been lying around for many an age...
> 
> To share a FT_Face object between multiple instances of that font face at
> differing sizes...
> 
> And behold did I marvel at the fact that the code worked. I was astounded that
> it was almost flawless and "just worked (tm)". And it was then I basked in the
> glory of the judicious use of a face cache and FT_Set_Char_Size().
> 
> But the gods did not want life to be this easy. They did gnash their teeth in
> chagrin and thusly so sent their evil demons to the earth to lay waste to such
> happiness. They did so on the day I ran my performance tests...
> 
> AAARGH!
> 
> (sorry about the over-dramatic prose).
> 
> So yes. I suspected FT_Set_Char_Size() would be a little nasty so I already
> wrapped it so it would only call it *IF* the font face that was shared was
> REALLY being used at a new size... even with this in, the performance hit was
> shocking! But now instead I have to load the font face multiple times instead
> if I use the same font at more than one size.
> 
> "Aaah but then just use a glyph and font metric cache yourself" I hear you
> say! well - I already have a glyph cache. It's fast. I'm happy. I could add in
> glyph metrics to it in a snap thus keeping calls to FT_Set_Char_Size() only
> when new glyphs are generated. This should solve most of the deathly ills,
> except there is a new little demon to deal with: KERNING! I would need a way
> of caching the kerning information. I want to keep hinting - it looks much
> better. I want kerning - it looks much better. I can keep my own cache of
> glyph and general font metrics I use... EXCEPT for kerning. I cannot see a way
> of getting kerning information out of freetype2 so I can make my own kerning
> table and thus not have to call FT_Set_Char_Size() unless generating a new
> glyph bit/pixmap.
> 
> Soooo... after a lot of windy longs... you guys know freetype much better than
> I- I'm asking... what do I do? How do I keep performance AND manage to share
> face outlines and save extra loads and memory? Is my only choice to turn off
> kerning? Is there something int the freetype api that lets me get some form of
> kerning rule table per glyph (ie left if glyph = X, kerning modifiers are for
> right glyph: Y = -2, Z = 1, A = 3, B = 5, G = 2, J = -2)... ?
> 
> It'd be really helpful to know what to do other than just shrug and go "either
> load the font multiple times... or ditch kerning" :(
> 
> BTW. Thanks for all the code. Nice yummy stuff. So happy it exists. Good work!
> :)
> 
> --
> ------------- Codito, ergo sum - "I code, therefore I am" --------------
> The Rasterman (Carsten Haitzler)    address@hidden
> 熊耳 - 車君 (数田)                  address@hidden
> Tokyo, Japan (東京 日本)
> 
> _______________________________________________
> Freetype mailing list
> address@hidden
> http://www.freetype.org/mailman/listinfo/freetype
> 
> 
> 
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    address@hidden
熊耳 - 車君 (数田)                  address@hidden
Tokyo, Japan (東京 日本)