When a given file is, for example, unreadable or unwritable, the
error message should use ^J instead of ^@ in its name.

In path names and file names, 0x0A means an embedded newline and
should be shown as ^J, but in anything related to the file's data,
0x0A is an encoded NUL and should be displayed as ^@.

So... switch mode at the two main entry points into the "file system"
(reading in a file, and writing out a file), and also when drawing the
titlebar.  Switch back to the default mode in the main loop.

This fixes https://savannah.gnu.org/bugs/?49893.

Decode 0x0A bytes to 0x00 when saving the search history, and encode
them again when reading the file back in, to prevent nano from hanging
or aborting when encountering 0x00 on a line by itself.

After updating the color palette, the corresponding new window content
should be drawn immediately, before some other part of the code calls
doupdate(), to prevent the old content being shown in the new colors.

This fixes https://savannah.gnu.org/bugs/?49912

.
Reported-by: Mike Frysinger <vapier@gentoo.org>

This fixes https://savannah.gnu.org/bugs/?49884.

Taking the terminating newline into account, and that there is
at least one digit per number.

A filename might contain spaces, so we can't look for the numbers
(the second and third elements) starting from the head of the line
-- we have to start at the tail and work backward.

This fixes https://savannah.gnu.org/bugs/?49879.

(I don't /see/ single-letter variables -- they are too small.)

It is irrelevant -- the line is discarded as soon as the data has been
extracted.

So they will not break a line in the positionlog file in two.

(Strangely, the reading in of such a log file already decodes
nulls back into newlines.)

This fixes https://savannah.gnu.org/bugs/?49877.

Because changing anything to a null effectively means to truncate
the name.

This fixes https://savannah.gnu.org/bugs/?49868
and fixes https://savannah.gnu.org/bugs/?49874.

Most full paths are needed only temporarily and will be freed within
milliseconds.  Only 'full_operating_dir' and 'backup_dir' continue to
exist for the whole current session.  Any partition, too, will soon be
unpartitioned, so the extra reallocation is just a waste of time.

It doesn't align anything -- any allocations are already aligned to
whatever multiple is required -- it just shrinks the allocated space.

And which normally is just some ten or twenty characters long, and
never gets wildly overallocated.

Nano would crash straight afterward if any of these asserts would fail,
so they don't add anything.  A few others are simply superfluous.

The functions read() and fwrite() take size_t, not ssize_t.

And line numbers in the file should be displayed as a long type instead
of an int, since the effective type of ssize_t is not int, but long.

This fixes https://savannah.gnu.org/bugs/?48622.

Achieve this by reusing the code that gives feedback when trying to
save a buffer while using --tempfile and the file has no name yet.

This fixes https://savannah.gnu.org/bugs/?48622.

Also improve or correct some comments.

If it would, the returned file descriptor would make nano crash,
because the corresponding stream has not been opened.  And when
returning zero instead (as the code did originally), nano would
open an empty buffer, although it claims to be reading the file.

In short: I think this is a leftover of an attempted fix of
https://savannah.gnu.org/bugs/?25297, from commit 2823c99e.

This fixes https://savannah.gnu.org/bugs/?49423.

Also remove some cluttering conditional compilation.

Instead of doing this in two places: after interpreting shortcuts,
and after injecting characters.