[Nano-devel] updates, replies, etc.

[David Lawrence Ramsey]

I know I've been quiet for awhile.  I've been busy again, though.  The
following changes are in CVS:

* the ^R^X^X and ^W^T^T toggles are now implemented, and resetstatuspos
is no longer global either, as described in the TODO list; the "Execute
Command" prompt also has a multibuffer toggle now to be consistent with
the "Read File" prompt

* with the new DOS/Mac format autodetection code, having command line
options for those formats doesn't make sense anymore, so they've been
removed; the toggles have also been reduced from global to local to the
"Write File" prompt, for consistency with Append/Prepend; finally, the
Mac format toggle has been changed to Meta-M, since it no longer
conflicts with the global mouse support toggle; the -D/--dos and
-M/--mac options are no longer in the nano.1 manpage or the nano info
page either

* the IEXTEN termios flag is now turned off, as it is in nano 1.2.x, in
terminal_init() (maybe this is the cause of the change in keypad
interpretation between 1.3.3 and 1.3.3-cvs that Vincent Raffensberger
noticed a few months ago?); terminal_init() is also called in all cases
where it's needed now

* current_x and mark_beginx are now size_t's, in accordance with DB's
patch, plus a few additions of mine (including a fix for a potential
infinite loop in do_prev_word() when current_x is unsigned)

* all my attempts at making marked internal spell checking and
replacement work properly have failed, although with each failed attempt
I've understood more of the code and hence have been able to make a few
cleanups; search_last_line is no longer global, the wholewords flag
should now work properly when regular expression searching is enabled,
the length of the match is now returned in a parameter since it has to
be calculated in order for the last item to work at all, and whether we
canceled is now returned in a parameter so that canceling in the middle
of the replace prompt when running the internal spell checker will now
cancel completely out of spell checking

* the original code to implement marked replace has been ripped out, as
it was hackish and didn't apply to findnextstr() as it should have; I
can properly implement marked spell checking and replacement at the
findnextstr() level by moving the code in write_marked() to set the
marked text up as the entire file into its own function and calling it
before and after findnextstr(); I'm going to get this done soon so it
can be in 1.3.5

* the cursor position is properly preserved when inserting a file into
the current buffer; it now takes placewewant into account

* a few help text updates

* a few c-file regex updates

Jordi's comments on UTF-8 support: good.  setlocale() is in both ANSI C
and POSIX.1, so it should be available on most decent systems.  I'm
looking into supporting UTF-8 without using the wide version of ncurses,
and I think I can do it, but it'll take longer, so I'll unfortunately
have to put it off one more time.  (Parts of DB's UTF-8 patch will still
be useful, too.)  There's preliminary support for getting and ungetting
UTF-8 characters in (un)?get_binput() now.  If the int returned is
outside ASCII range and func_key is TRUE, it's an extended keypad value,
but if func_key is FALSE, it's a wide character value.

Jordi's suggestion to add the latest glib-2.0.m4 file: applied.  I added
the one from glib 2.4.7.

Mike Frysinger's suggestion to add an option to use the always-blank
second line of the screen: good, but not for 1.3.5, as it'd require a
lot of reworking of those parts of the code that assume that topwin is
always 2 lines (the titlebar and the blank line).

DB's fix for a potential crash when unlink()'s parameter is NULL:
applied.  It should make nano more portable, especially when glibc isn't
being used.

Rocco's fix for the bug in the internal spell checker: applied.  
Apparently the internal spell checker isn't widely used if this wasn't
caught until now, but it's good that it was.

Chris' suggestion to use F16 (Shift-F4) as an alternative key for "Where
Is Next": good.  If F16 is going to be used for something, though,
shouldn't F15 be used too for consistency?  (Maybe F15 can be an
alternative key for the mark [Ctrl-^/Meta-A]?)  The Shift-(function key)
display in the help menu is a bit problematic, though.  It'll work if
there are no function keys past F12, but it probably won't if there are
function keys past F12.  I can add the F15 and F16 key support in the
meantime, though, if no one else objects.

After the marked spell check and replace code is done, I'll update the
version numbers and modification dates in the documentation and there
can be a release candidate.