As a small service to the user, reject empty regex strings,
because an entirely empty regex simply doesn't make sense.
Inspired-by: Elia Geretto <elia.f.geretto@gmail.com>

When the marked region covers only a single line (or a part of it),
its new endpoint is not simply the length of the last line of the
spell-checked text, but instead the old endpoint plus the /change/
in length.

This fixes https://savannah.gnu.org/bugs/?50316

.
Reported-by: David Lawrence Ramsey <pooka109@gmail.com>

The inversion of the logic in commit a6605dce was incomplete -- oops.

This fixes https://savannah.gnu.org/bugs/?50306.

The previous code only directly refreshed the screen when the margin
changed due to toggling line numbering on.  When the margin changed
due to toggling it off, it would indirectly refresh via do_toggle().

When replacements are made, nothing needs to be reset any more
(it was done insufficiently anyway).  Just make sure the screen
is refreshed when all is done -- this may be superfluous when
doing interactive replacements, but not when replacing all.

The lines that come after an unpaired start have to know about this.

This fixes https://savannah.gnu.org/bugs/?50293.

When coloring a line, look only at the multidata of the preceding
line, and based on that determine what to seek in the current line.

This fixes https://savannah.gnu.org/bugs/?50292.

When painting a line, the multidata of the line /before/ it is valid
in most cases: it was determined just a moment ago.  And it tells us
all we need to know: whether there is an unpaired start match before
the current line or not.

The only exception is when painting the first line of the screen:
the multidata of the line before it might be stale.  So for the
first screen line we will always have to do some backtracking.
But that is left for later.

This fixes https://savannah.gnu.org/bugs/?50121.

These two functions unnecessarily copy make_new_node(); use that instead.

If we're somewhere deep into the file and do a spell check, and the
first misspelled word happens to be right there, onscreen already,
then this word does not need to be centered -- it /should/ not be
centered.  We should scroll only when necessary.

Just like when spell-checking, there is no need to wrap around when
doing replacements in a region, because a region cannot straddle the
buffer's edges.

This is the proper fix for https://savannah.gnu.org/bugs/?50158,
and fixes https://savannah.gnu.org/bugs/?50273.

Scrolling works on rows, not on lines.  Accordingly, rename the nlines
variable to nrows.  And rows should be of the type int, not ssize_t.

Make it clearer we're comparing the number of rows, not string lengths.

After the changes to the search routine, it is no longer necessary to
take one step back before starting a replacement session.

This fixes https://savannah.gnu.org/bugs/?50147.

Even when a match falls within the marked region, this does not mean
that it is a true match when already the whole file has been searched,
because then this is the second time we find this match.

This fixes https://savannah.gnu.org/bugs/?50158.

The spell fixer does not provide a beginning line, so the search routine
should then not try to refer to any data of this line.  Also, since the
changes to the search routine there is no need any more to retreat one
step before starting to search for a misspelled word.

This fixes https://savannah.gnu.org/bugs/?50159.

When the user answered No to a replacement prompt, skip the match
at the current position, so we don't stay stuck there.

Also, when replacing in the backward direction, or the match is
of length zero, skip the match at the current position, to not
get stuck.

This fixes https://savannah.gnu.org/bugs/?50126

.
Reported-by: David Lawrence Ramsey <pooka109@gmail.com>

This also fixes https://savannah.gnu.org/bugs/?50137,
and refixes https://savannah.gnu.org/bugs/?48635,
and fixes https://savannah.gnu.org/bugs/?50144.

That is: remove the special treatment of BOW anchors, and instead make
regexes match against the whole line instead of against an artificially
shortened one, because the latter method creates ghost matches: matches
at the starting point of the search that aren't really matches when seen
in the context of the whole line.

This fixes https://savannah.gnu.org/bugs/?50030.

When the tail of a match falls outside of the marked region,
it is in fact not a match and should not be replaced.

This fixes https://savannah.gnu.org/bugs/?50136.

When a replacement happens right at where the cursor sits, the position
of the cursor should not be adjusted, because the real cursor position
is between the current character and the preceding one: the place where
the cursor is shown is in fact right /after/ the insertion point.

This fixes https://savannah.gnu.org/bugs/?50134,
and fixes https://savannah.gnu.org/bugs/?50135.

There is no need to compute the line length: just avoid overstepping
the terminating NUL byte when being forced to advance the index.