Re: [g-i] Stripping code in ttf-freefont, do we still need it?
I'll try to concentrate on this message all the answers to all the
previous messages on this thread
> Eddy, Davide, are these scripts still useful or are they redundant
> with font stripping that may be planned in the g-i build system? They
> hardcode the kept (or excluded?) ranges which sounds not really
> scalable to me.
the infrastucture for stripping the fonts (script + file with ranges)
will be needed; currently is not excactly defined the set of ranges.
If it is true that dejavu will be used as default font, I have a rough
idea of how the freefont range should look like.
Eddy's idea about the debhelper script sounds interesting but I think
there's something wrong in the idea about using debs without the need
of udebs. As Frans said before, we should keep font stripping away
from the build system, mainly because IMO it's too heavy (in terms of
cpu cycles) and if we decided to make it lighter we would depend on
results computed somewhere else which would make things quite
complicate / delicate.
> Davide, we should have a discussion related to this issue. How can we
> do font stripping based _exactly_ on used the ranges in the D-I
> translations and make this for all used fonts? It would be nice to
Current scripts are making things easier, but I wouldn't rush too much
on this: once we are sure we cover all languages (first goal) my idea
is to remove the glyphs not used by any font file (safe option): this
means about ~1200 glyphs that could mean ~500 Kb. After this we can go
even more into details but anyway I don't think we would save too much.
To give more sense to the numbers above, we're using about 2.3 Mb for
fonts and I think we can reach ~1.8Mb which I think is not bad at all (less
10% of the space needed by g-i)
Tiscali ADSL 4 Mega Flat
Naviga senza limiti con l'unica Adsl a 4 Mega di velocità a soli 19,95 € al mese!
Attivala subito e hai GRATIS 2 MESI e l'ATTIVAZIONE.