conversion tools currently being constructed.

Add a chapter from Jim Fulton on using Misc/Makefile.pre.in.  Still
preliminary.  The "Dynamic Loading" chapter needs to be updated (and
possibly removed, since it's no longer an issue for most (any?)
users.

the previous patch wasn't -- there was a path through the code that
bypassed all initializations.  Thanks to Just for reporting the bug!

different internal indentation levels.

than integers and recommending sys.exit("message").
(I see this as Python's answer to Perls ``die "message";''.)

LaTeX-based ESIS streams to be a little better structured, and generally
perform clean-up.

Preliminary.

Be more informative in an exception message.

Don't duplicate the information on what's empty; rely on the input
data for that.  (This means that the DOM may need more work.)

Make the table of attribute information contain only one kind of
entry.  Simplifies some stuff in subconvert().

semantics between CPython/JPython when the script source is piped to
stdin.

probably only usable for the Python docs.  But also probably more
useful than latex2esis.py.

probably only usable for the Python docs.

do_delete(); do_clear() was meant.  By Greg Ward.

correct Python shell (CPython or JPython).

Because it might be a common mistake to pass a single string, this
situation is treated separately.

Since we were making a copy of the longopts list anyway, we now use
the list() function -- this made it necessary to change all uses of
the local variable (and argument) 'list' to something more meaningful,
i.e., 'opts'.

Also added docstrings (copied from the library manual) and removed the
(now redundant) module comments.

running in a non-threaded environment.  He added some #ifdefs that fix
this.

Noted by Stefan Witzel.

Noted by Stefan Witzel.

Tip by Vladimir Marangozov.

METH_VARARGS conventions and PyArg_ParseTuple(), and document the flag
and where to look for PyArg_ParseTuple() info.

Response to comment from Don Bashford <bashford@scripps.edu>.

<dalke@bioreason.com>).