UNIX file extensions

Mike Don't have strength to leave Meyer mwm at eris.BERKELEY.EDU
Sat Dec 13 17:31:51 AEST 1986


In article <1694 at cwruecmp.UUCP> cwruacm at cwruecmp.UUCP (Kronen Insultants) writes:
>
>I don't know if these are really enforced anywhere, but how about
>
>	.el (or .ell) Lisp (probably Franz) source file

I think Franz uses .l (been a long time since I looked). GNU Emacs
uses .el for Elisp and .elc for compiled Elisp. It "enforces" this by
looking for <name>.el and <name>.elc when asked to load <name>.

Similarly, Gosling/Unipress Emacs uses .ml for Mock Lisp files.

For yet more, Scheme uses .scm and CLU uses .clu. I think the Arizona
Little SmallTalk uses .st. Icon uses .icn.

But for Unix, the point is kind of moot. The only "enforcement" is
what certain application programs do with them. I've compiled
pseudo-devices by symlinking them to "gort.c" in the current
directory.

Since such tricks will always work, and nothing in the OS enforces
those typings (thank GOD!), what's the point of worrying about what
the extensions are? Use the contents. Compress checks for them, and a
file system browser I wrote back in the v6 days did the same kind of
thing. Much more reliable.

After all, EVERYONE knows that .s files are SLOGO source, right?

	<mike



More information about the Comp.unix.questions mailing list