Also contains getuser(), which returns the username (not prompting though).
These work on Unix and Windows!

Consistently use ' instead of ":  this conforms to my unofficial stance that
constants should be presented using the repr() whenever it makes sense.  This
isn't because I think repr() is the greatest thing since spam, just that it's
an easy to adopt consistency standard.  (It also holds no new surprises!)

do_cmd_lineiii():  Fix markup style conversion to HTML tags for first column.
	(The endtag wasn't formed properly.)

\seemodule.  This is far easier than figuring out the real problem, but it
works.

library have no relationship with GNU software other than implementing some of
the same algorithms and file format.  (The zlib module is more general.)

Clarified text on 8-bit cleanness, complemented sets like [^5], and {m,n}
  qualifier.

is the full release version name ("1.5.1beta1", for instance).

Add section for previously documented obsolete modules.  Includes ni, rand,
and soundex.

fall back to using os.curdir instead; if it is fine, don't use
os.curdir at all.

lines to doc strings.

is the full release version name ("1.5.1beta1", for instance).

(Adapted by Just, I believe.)

Basically, it now works and produces a nice-looking HTML index.

Shows some promise for extensibility, but still young.  Needs useful comments.

left to the buildindex.py script.

Change the field separator used in the index.dat file so utilities like grep
are useful on these files; they can't typically handle NUL bytes.

Protect index entries for environment variables from being handled as
case-insensitive by buildindex.py.

looking really bad in popular browsers, mostly due to vertical size differences
between the roman and monospaced fonts.

This is the last separation; each module now gets its own section.  This will
be useful in performing conversions.