Add a -q option for "quiet" operation. Nothing gets out (not even stderr,
since that's tossed to stdout for capture in the .how file). It's a good
idea to use -l to get the log in case there are problems. Maybe this should
be the default and add a --verbose option to get the stdout? There's way
too much junk output.
Kerberos from Python" HOWTO (this will change shortly, I hope).
This needs to be run from it's location in the Doc/tools directory, and
not via a symlink. I'll try and fix this in the future.
The help message needs work.
- the .cvsignore files; these are only useful in the CVS tree
- ref/ref.pdf
- ref/ref.ps
After compression, these last two account for about one-half MB, and half
the tarball. They can be picked up from the appropriate format dist if
needed.
Chunk.__repr__(), main(): Allow Chunk objects to get the buffer, to make the
representation contain text instead of a pair of indexes. This makes debugging
a little easier.
messy, but the thing seems to be working without bombing completely today.
Formatting lib.texi with TeX doesn't seem to do too badly, either!
Info formatting isn't quite there; that might just have to disappear this
time.
"""Combine similar index entries into an entry and subentries.
For example:
\item {foobar} (in module flotz), 23
\item {foobar} (in module whackit), 4323
becomes
\item {foobar}
\subitem in module flotz, 23
\subitem in module whackit, 4323
Note that an item which matches the format of a collapsable item but which
isn't part of a group of similar items is not modified.
"""
This results in a much more readable index, with less repitition of text;
especially for common method names.
Lots of support for new macros defined in myformat.sty; including the new
indexing macros, seealso environment & friends, and the byte code instruction
support.
navigation links for HTML 3 version.
Forced a blank line above the footnotes separator for HTML 2; at
least one page did not get this spaced correctly.
print section titles even when the debugging output is not enabled.
Added -3 option to generate HTML 3.0 constructs where meaningful.
Removed repititive garbage generation: the old version added simple
descriptive comments after every datadesc/funcdesc/*desc entry:
function(args) -- function of module xxxx
Description....
These comments are no longer generated:
function(args)
Description....
to 4 spaces per level (no longer 8).
(Makefile): Use .pyc versions of partparse.py and texi2html.py to generate
converted documentation formats. This reduces the startup costs;
probably doesn't affect anyone but me in reality, but helps when
working on the docs.
* Makefile: added 'ext' to 'all' target
* ext.tex: more changes towards a readable text
* lib4.tex (posix): added set{uid,gid}
* lib2.tex (array): restored doc for typecode and itemsize (which were
there but not visible for dir())
(formerly ../misc/{EXTENDING,REFCNT,EMBEDDING}). Also affects Makefile.
* text2latex.py: script to do part of the conversion from an plain ASCI
text file (in my particular style) to LaTeX.
(Chapter/section/subsection headers, and verbatim sections.)
* partparse.py, texipre.dat, fix.el, Makefile: Minor cleanup of latex ->
info conversion process (at least it works again, and with less
debugging output). Removed fix.sh.
* lib1.tex (section{Built-in Functions}): adapt description of str() and
repr() to new situation.
* lib3.tex (Module os): added exec*() variants.
* lib3.tex (Module posix): added execve().
* lib2.tex (Module array): documented reality; remove typecode and
itemsize, add byteswap, rename read/write to fromfile/tofile, and
re-alphabetized.
* lib1.tex (Built-in Functions): renamed bagof() to filter().
Fred Drake