Re: [Devel] initial release of FTLayout/GXLayout

[Antoine Leca]

Hi Owen,

Owen Taylor wrote:
>
> In the middle-term, I'd certainly be happy if we can accomplish our
> project of moving the OpenType table parsing portions of Pango back
> under the FreeType umbrella, which would simplify things here.

Thanks Owen, and David, for the detailled explanations.
So the situation is even less clear than I thought... :-(

Regarding Owen's "conclusion", and taking in account Masatake's work, it
boilds down we have a real problem with ot-unicode.[ch], which in fact is a
rewrite/extract/"modified" version of GLib, and as such falls under LGPL
(IIUC), so cannot be easily "retrofitted" into the dual license plan of
Freetype.
However, this seems a fairly simple API (reading the .h) provided here, so
it appears this could be "reinvented" without actually using the GLib
source, for example by writing some (Python?) script to extract the data
from some Unicode data file. Correct?

(Please, don't exhibit any such script extracted from GLib: we cannot use
it, since it would be LGPL or GPL covered, so difficult if not impossible to
include in any FTL-covered package; furthermore, I have not seen it for the
moment, and I prefer to stay "virgin" of this, which will allow me to write
a clean-room implementation).


I take for granted, given Owen's post, that the rest of Owen code can be so
retrofitted. As David pointed out about copyright holders, this should be
done "by Owen" (really, by Owen on behalf of Red Hat), or at least with
clear disclaimer from him. To have it be done by Masatake is not clear to me
to be correct, for instance (see David's example about "waste").

Another point is that on its own work, Masatake did not assign a shared
copyright to David and Werner (in addition to him or RedHat), so perhaps
this is a problem, at least to host this code in the Freetype CVS, and
moreover to release it. David?


Antoine