[Savannah-hackers] Re: FOO.gnu.org

[Ron Peterson]

On Sun, 9 Dec 2001, Ron Peterson wrote:

> On Thu, 6 Dec 2001, Richard Stallman wrote:
>
> > Someone suggested in October that we could establish FOO.gnu.org
> > as a virtual host alias for http://www.gnu.org/software/FOO
> > for all supported values of FOO.  I worried at first that gcc.gnu.org
> > might be a stumbling block, but apparently it will not be.
> > We just have to leave the current definition of that host unchanged.
> >
> > So, do people want to make this change?
> >
> > Does anyone see any reasons we should not do this?
> > Any problems it would cause?
>
> Is the "for all supported values of FOO" a problem?
>
> http://www.w3.org/Addressing/URL/5_BNF.html describes valid URL schemes.
>
> Of course, any project that has a /software/FOO entry now has a valid FOO.
> Maybe the question is, do we care about establishing any kind of
> consistency for (virtual) hostnames?  E.G. - should a project named "GNU
> FOO" be called "gnufoo" or "foo"?  Technically, it doesn't matter.

Also, the virtual host alias must be a valid hostname.  I.E. -
underscores, the '+' symbol, and other valid URL characters can't be used
in a hostname.  E.G. - www.gnu.org/software/c++ is o.k., but c++.gnu.org
is not.

I'm sure others know better than I, but I understand valid hostnames are
alphanumeric, and can include a minus sign or a period, when the period is
used to delimit components of a domain style name.  The first character
must be an alpha character.  The last character must not be a minus sign
or period.

-- 

-Ron-
https://www.yellowbank.com/