[Ecls-list] Critical change

[Erik Huelsmann]

On 1/18/07, Juan Jose Garcia-Ripoll <jjgarcia at users.sourceforge.net> wrote:
> 2007/1/18, Dean O'Connor <dean.oconnor at ite.com.au>:
> > Perhaps it would be a good idea, to prefix *all* exported/global
> > functions/vars with "ecl_" even if it means "ecl_cl_" and "ecl_clos_" etc.
> > A hassle perhaps, but then there's only 1 prefix (~namespace) to worry
> > about.
>
> I do not know. It is not a difficult change, but it will make much of
> the code almost illegible. Also, some of the names have probably close
> to 32 characters, which is the limit of some compilers around. I
> wonder whether we will have problems any time soon.

There are compilers with that limit? I work on the Subversion project
and we prefix all symbols with 'svn_(subcomponent)_' where
subcomponent can count upto 7 characters ('ra_dav_'). We've never
heard of a limit like that (and as a matter of fact, we're exceeding
it in a lot of places!).

Yes, it is sometimes troublesome to read the sources, but it's nice to
be sure to know symbol collisions will be extremely rare.

HTH,

Erik.