[Texmacs-dev] Re: C++'s frustrations

[Henri Lesourd]

Abdelrazak Younes wrote:

> Henri Lesourd wrote:
>
> >  Probably under Linux it works most of the time,
>
>
> _always_
>
> >  and under windows
> >  it's more often broken:
>
>
> MSVC is really the defacto standard (KDE, python, LyX, etc all use 
> it). I don't see the problem here. In any case, I think, but I am not 
> sure, Borland and Intel compilers are compatible while Mingw's g++-3.2 
> is not

Thus there is a problem: we need to go into the
business of telling plugins implementors which
tool to use.


> (not sure about recent version of g++). So yes, I think that choosing 
> either MSVC or Mingw would be a hard requirement for plugin  
> developer. Or you can just see those two compilers as two different 
> platforms and deliver a TeXMacs binary for both. That is exactly what 
> the KDE on Windows people do by the way.
Interesting information.


> >  Or perhaps do you have some other interesting new information about
> >  C++ which could cure my doubts ?
>
>
> I hope my argument above did so. I don't think that imposing a 
> compiler for plugin development is a very hard requirement,

The problem is that in such a situation, ultimately
it's the *final user* which will have to deal with
these issues: he will have a TeXmacs installed,
download a plugin, and then he will need to be
able to figure out that this compiler compatibility
problem *exists* (first), and then figure out how
he can obtain the address@hidden related info...


> especially nowadays where you can have free (as in beer) compiler on 
> all platforms.
Free is not the only criterion: not forcing users to
think about issues they should never have to enter in
the first place is another important point.

It's the very point M$ understood perfectly, and
that OSS operating systems repeatedly missed since
the very beginning, by the way...