Re: [Devel] Ready for 2.0.8

[David Turner]

Hi Werner,

Werner LEMBERG a écrit :
> 
> > I've been testing it and have found no problems with the 9000+ fonts
> > on my machine.  It's easy enough to test; walk the encoding linearly
> > and compare with walking via Get_Next_Char.  I'd appreciate having
> > the API available in 2.0.8 as it turns a 10 minute task into a 2
> > second task and I'm nearly set to release the dependent libraries.
> 
> I think that the code is quite reliable, so I don't mind having this
> new API.  Note that we have to increase the version of the DLLs for
> all platforms in that case.
>

Actually, I simply want to avoid the situation where we discover three
hours after the release that the API is not sufficient for what Keith
(or other people) wants, and that another exotic function is needed..
A few days later, we find out that it's still isn't sufficient, and
that yet another API is needed, etc..

In the end, we'll have several related APIs to support till the end
of time, while we could have chosen only once from the start by being
a bit more careful..

This model of "progressive additions" of APIs to the library is what
led us to the internal mess that is now within the sources and that
I'd like to get rid of.. For example, I have made changes to the 2.1
branch that are not compatible with Keith's latest patch (which was
included since it has a higher priority in my opinion), so I've started
them again.. (that's not much of a problem actually..)

Numerous little API additions are generally not compatible with major
internal re-workings. Another solution is to by-pass completely 2.1 and
start working on 3.x directly, so that any API can be crammed in the
2.0.x branch "at will", using the time differential between the two
versions as a test bed for new features..


Regarding the specific FT_Get_Next_Char API, it seems to be nice
except that:

 A - I'd prefer having a "FT_Get_First_Char" function, instead of
     calling FT_Get_Next_Char(0) to retrieve the first character
     code in a charmap..

     the latter assumes that '0' is never a valid character code,
     and I'm just too skeptical to believe that this will never
     be a problem, even if we can't see one now


  B - I'd prefer that the "FT_Get_First_Char" and "FT_Get_Next_Char"
      functions be able to return the glyph index corresponding to
      the character code as well, like:

            FT_ULong   FT_Get_First_Char( FT_Face   face,
                                          FT_UInt  *aglyph_index );  // optional

            FT_ULong   FT_Get_Next_Char( FT_Face    face,
                                         FT_ULong   char_code,
                                         FT_UInt   *aglyph_index );  // optional


'A' can be implemented right now (calling FT_Get_Next_Char(0)
until we properly implement it.. On the other hand, 'B' needs
reworking the latest additions..


That's why I'd prefer to release 2.0.8 right now without the additional
API, even if this means we're going to release 2.0.9 a few days after
that. I simply value critical bug-fixes at higher priorities than
additional features for now..


If you don't object, I'll make the release today..

Cheers,

- David