libcdio-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Libcdio-devel] RFC: Savannah issue bug #35745


From: Rocky Bernstein
Subject: Re: [Libcdio-devel] RFC: Savannah issue bug #35745
Date: Tue, 13 Mar 2012 10:05:18 -0400

On Mon, Mar 12, 2012 at 5:09 AM, Nicolas Boullis <address@hidden>wrote:

> On Sun, Mar 11, 2012 at 11:09:35PM -0400, Rocky Bernstein wrote:
> > A patch was recently submitted: https://savannah.gnu.org/bugs/?35745
> >
> > Any comments on this?
>
> I had no idea that identifiers that begin with _[_A-Z] were reserved,
> but if they really are, I think it is sane to replace them, at least for
> macros use as "header guards", which, I guess, nobody uses outside the
> header files themselves.
>

As best as I can tell, the argument goes that various compilation systems
and system libraries are free to start with _,  the suggestion is in user C
code to avoid starting with _ altogether lest there be a conflict. In
reality the likelihood of a compilation system or system library
conflicting with one of the names used in the libcdio headers is pretty
small. Furthermore, since the names in question are preprocessor names,
those names are gone by the time linking takes place; this reduces further
the chance of a real (as opposed to imagined) problem.

People sometimes take a real concern and then frame rules around it,
greatly expanding the margin of error in the process. Then others start at
this point taking the rule this as religion and perhaps the margins even
grow more. Lost sight is the original concern.

I think it okay and worthwhile to avoid chance of conflict. However at some
point the likelihood gets so vanishingly small that it is no longer worth
worrying about.  And it shouldn't be at the expense of making programs ugly
and less user friendly.  But this is a matter of taste.


> As for the patch itself, it adds a 32-character random-like suffix to
> all identifiers. I think this is really ugly, and I can see no good
> reason for this...
>

I agree. I think the intent was to avoid arbitrary clash. But where does
one stop at? Would one really be certain at 64 characters?  But all of this
is in my opinion this is misguided because the names themselves are
generally already long enough.

But thanks, I just wanted to get independent views.

>
>
> Cheers,
>
> --
> Nicolas
>
>


reply via email to

[Prev in Thread] Current Thread [Next in Thread]