Syncing up the MIME databases
jrb at redhat.com
Wed Nov 5 20:13:35 EET 2003
Thomas Leonard <tal00r at ecs.soton.ac.uk> writes:
> Originally by a script, but it should be done by hand now since there have
> been changes made to it since then.
Oh yuck. What about moving to an intltool-like setup so we can at least
move the translations out. Putting the translations in the file makes
you vulnerable to the first person who edits the file with a non-UTF8
safe editor. It would be a lot less cumbersome to deal with, too.
> > There were a few MIME-types that caught my eye as being potentially
> > problematic. First, the inode/* MIME-types need to be 'x-' escaped, if
> > nothing else. I can't find anything at the IANA that indicates the
> > presence of a inode/ content type.
> There are lots of should-be-x-escaped ones. The policy was not to change
> anything if lots of people were using them already. For the inode/* stuff
> we should be OK. I think it unlikely that IANA will create an 'inode'
> group, but even if they do, we'd be exceptionally unlucky if they then
> created a subtype 'directory' with a different meaning ;-)
That doesn't mean we can just stomp all over a global namespace. We
shouldn't go around creating our own toplevel content-type.
> The general agreement, as I recall, was that we should be using vendor
> tags for new types (not x- prefixes), and keep the ones we were already
> using the same.
Makes sense for some types, but not all.
More information about the xdg