On Fri, 4 Feb 2000, Allan Rae wrote:
> On Fri, 4 Feb 2000, Kevin Atkinson wrote:
>
> > On 3 Feb 2000, Lars Gullik Bjønnes wrote:
>
> > > C is not good for C++ programs. (usable yes, but not good)
> >
> > Do the other developers fell that way?
>
> I've already said you might be better off wrapping ispell with your aspell
> C++ interface and then once that works worry about a C wrapper.
My Aspell C++ interface using exceptions, and templates. If you can
compile and use my interface you can compile Aspell so wrapping Ispell in
aspell will almost be pointless.
I could develop a new interface which conforms to the mozilla portability
guidelines (http://www.mozilla.org/hacking/portable-cpp.html). However
that still leaves out C programs from using my interface. Then I could
wrap wrap my C++ interface in C.
Or I could develop a C interface to Aspell and Ispell. I can then wrap
this in C++.
Do all the LyX developers think I should do the second? If you want to
kiss gcc 2.7.2 good by then perhaps you MIGHT be better off using aspell
directly. I don't plan to make my Aspell/Ispell interface library as
functional as directly using Aspell. However, the Aspell/Ispell interface
will provide all the needed functionality for most applications. The
stuff that won't be included is low level Aspell specific stuff like
converting words to there sounds like equivalent, creating new word list
classes, and the like.
Please look at the CVS source to see what my current interface
looks like as it changed a bit from aspell .28.3. The interface in "lib"
is mostly stable except for the tokinizer and check functions. The stuff
in "filter" is extremely unstable as it is very new and has many problems
I need to address. The stuff in "util" is independent classes and
functions which are also very unstable unless they are visible by header
files in "lib/inc". The "src" directly is the actual aspell utility.
PS: Is there a way to get LyX to read in a .lyx file and generate a .tex
file from the command line so that I can fully automate the building of
the manual.
---
Kevin Atkinson
[EMAIL PROTECTED]
http://metalab.unc.edu/kevina/
