The old_current line needs to be redrawn only if it differs from current,
and if it wasn't drawn already by the iteration for when the mark is on.

Also make the conditions involving horizontal scrolling more precise.

Instead of saving the current value of placewewant, then setting the
new value, and then passing the old value to edit_redraw() in seven
different places, just let edit_redraw() do this saving and setting.

In the bargain placewewant is now only recalculated when it matters
-- when allow_update is TRUE -- and not when it's superfluous.

Nano doesn't start doing anything with the edit window or the keyboard
until all files have been read in or a blank buffer has been opened, so
the case of openfile->current == NULL will never occur.

Also correct the comment -- because with multibyte characters, it is
very well possible that the screen column corresponding to current_x
is smaller than current_x itself.

Instead of allocating a fixed amount of 128 bytes, which will overflow
and segfault, adjust the allocation to the length of the file name, and
if necessary trim the file name to make the prompt fit on the screen.

This fixes https://savannah.gnu.org/bugs/?47511

.
Reported-by: Aapo Rantalainen <aapo.rantalainen@gmail.com>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

Now that findnextstr() no longer sets placewewant, we can can make
a copy of the old value just where needed: when a bracket is found.

In the innermost search loop, don't set placewewant, because this loop
is also used for replacing and spell fixing, when we don't really want
to be there: we are just passing through.  Not setting placewewant means
we don't need to save and restore it in those passing-through routines.

The value of placewewant is only relevant when doing cursor movement,
which doesn't happen during replacing nor spell checking, so there is
no need to keep placewewant up to date -- it is set when it matters:
at the end of go_looking().

Stop keeping track of the vertical screen position when searching for
something.  If nothing is found, current_y doesn't change, and all the
incrementing/decrementing was a waste of time.  If something is found
and it is onscreen, it is easy to calculate the new current_y.  And if
something is found and it is offscreen, then current_y is irrelevant,
because we will be either centering the found occurrence (searching)
or putting it on the top or bottom line (bracket matching).

(The above does not take softwrapping into account, but neither did
the old code, so this doesn't introduce any new bugs.)

(Also, when the search wraps, and the viewport is away from head or
tail of the file, and the found occurrence is within the viewport,
then the incremented/decremented current_y would be way wrong, but
this didn't have any adverse effects as far as I could tell.  It
seems that current_y is irrelevant in most cases.)

Add a third mode of scrolling, FLOWING, besides CENTERING and STATIONARY.
This is used for word and paragraph jumping (and for bracket matching,
but that worked correctly already), and only when focusing is FALSE.

The new mode prevents the screen from scrolling too many lines when
there are several blank lines at the bottom of the edit window and
the next word or paragraph is out of view.

This fixes https://savannah.gnu.org/bugs/?47194.

On iTerm2 on OS X, the Option+Arrow keys produce special sequences
that start with two escapes.  Catch these sequences and interpret
them appropriately as WordLeft / WordRight / Home / End.
Signed-off-by: Mike Scalora <mike@scalora.org>
Signed-off-by: Thomas Rosenau <thomasr@fantasymail.de>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

So it will fit on the screen when using a standard 80-column terminal.

Don't call the latter ones Escape sequences, because both Control
and Meta sequences can be entered using the Escape key.

This fixes https://savannah.gnu.org/bugs/?47544,
and thus also http://gnats.netbsd.org/51010

.
Reported-by: Adrian Siekierka <asiekierka@gmail.com>
Reported-by: Matthew Hall <mhall@mhcomputing.net>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

SVN revision 5748 could cause some 'else's to be orphaned.

This fixes https://savannah.gnu.org/bugs/?47610

.
Reported-by: Thomas Rosenau <thomasr@fantasymail.de>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

When finding a misspelled word, the length of the match is simply the
length of that word, and its span in columns is simply the number of
columns that it occupies.  Compute it thus directly.
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

When searching for a misspelled word takes a while, and the user
stops this search with ^C, then abort the spelling-checking session.
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5782 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

for whole words only, so it will be a whole word.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5781 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5780 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5779 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5778 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

for aptness, for contrast.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5777 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

that is taking too long.  This fixes Savannah bug #47439.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5776 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5775 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

whole spell-fixing session.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5774 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

(The measurable effect (during long searches, for example) is zero, though.)


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5773 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

an operating directory is not the same as restricted mode.

(In restricted mode, the file browser is not available at all.)


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5772 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

invalid escape sequence, and when entering a verbatim keystroke.
But leaving the cursor off during Unicode input, for extra feedback.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5771 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5770 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5769 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5768 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5767 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

anyway, and it will be placed back when it's needed: in the main loop.

This prevents a segfault laid bare by r5763: when trying, at startup,
to open a directory, there really is no open buffer yet.


git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@5766 35c25a1d-7b9e-4130-9fde-d3aeb78583b8