Re: operations on FILE streams

[Paul Eggert]

Bruno Haible <address@hidden> writes:

>   - Standardizing functions starting with '__' would encourage users to
>     adopt similar naming conventions (since users often use conventions
>     prototyped by the OS vendors)

I'm not that worried about it.  I think users will know what they're
getting into.

Standards can standardize names that are reserved to the
implementation.  This doesn't happen often, but it does happen, e.g.,
__bool_true_false_are_defined (for C99), _Exit (for Posix).

The other points you mention are about whether the existing interface
could be improved.  These are good points but this is a separate issue
from the naming convention.

> How can anyone expect users to understand that '__' symbols are reserved,
> if glibc documents such functions?

I think that message is pretty clear already -- or at least it's as
clear as it's ever going to be -- and this is true regardless of the
fact that the standard and some implementations document some __
symbols.

For what it's worth, Solaris and glibc both document these functions.

If it had been me, I would have omitted the leading __ from all the
names.  But the damage is done now, I'm afraid.  Existing code in
coreutils, Emacs, and I assume elsewhere already uses the __ names.
So the horse left the barn on this issue several years ago.

Anyway, it's not that big a deal.