When a fragment of help text does not contain any whitespace that
we can wrap at, just advance the pointer, because returning zero
would mean we stay forever at the same place: nano would hang.
That is, call display_string() just on the filename, not on the rest
of the prompt text.
This fixes https://savannah.gnu.org/bugs/?52967.
Reported-by: Andreas Schamanek <schamane@fam.tuwien.ac.at>
Instead of assuming that the entire contents of the keybuffer are an
escape sequence (even if this is far longer than any escape sequence
can be), remove from the keybuffer only those integers that actually
constituted a sequence, valid or invalid.
This fixes https://savannah.gnu.org/bugs/?48659.
Trimming trailing spaces is good, but we should not trim the space
(or tab or other blank) that the user just typed and that caused the
hard-wrapping to occur.
This fixes https://savannah.gnu.org/bugs/?52948.
Reported-by: Andreas Schamanek <schamane@fam.tuwien.ac.at>
The verbosity "hid" the really relevant information:
filename, line number, and the error message.
Valid values and other details can be found in the manual pages.
When the palette is getting initialized, it is too late to send any
error messages about the rcfile options to standard error.
This fixes https://savannah.gnu.org/bugs/?52871.
Reported-by: Brand Huntsman <alpha@qzx.com>
Signed-off-by: Brand Huntsman <alpha@qzx.com>
When we're just searching, we do want to report the occurrence at
the cursor (after coming full circle) as a match, so that we can
mention that this is the only occurrence. When replacing, we do
not want to replace this match twice.
This fixes https://savannah.gnu.org/bugs/?52888.
There shouldn't be two separate variables (but with the same name) that
count the number of digits seen. Fuse them into a single static var.
This fixes https://savannah.gnu.org/bugs/?52863.
Each leading tab is converted to two tabs, and any leading four spaces
is converted to one tab. The intended tab size (for keeping most lines
within 80 columns) is now four.
When unindenting/uncommenting affected the bottom line of the marked
region, keep affecting this line also during subsequent consecutive
indenting/commenting.
This fixes https://savannah.gnu.org/bugs/?52718.
If the marked region ends at the start of a line, do not include that
line in the indenting/undenting or commenting/uncommenting. This is
closer to what the eye would expect.
Unset the "Modified" marker only at the point where the file was last
saved -- if there is such a point, because it can be missing when the
undo stack was discarded.
This fixes https://savannah.gnu.org/bugs/?52689.
Reported-by: Liu Hao <lh_mouse@126.com>
Original-idea-by: Brand Huntsman <alpha@qzx.com>
The function get_keycode() should never return zero -- it should
return either the code from ncurses or the given fallback code.
This fixes https://savannah.gnu.org/bugs/?52682.
When the mark is on, instead of letting a <Tab> simply insert a Tab
character at the cursor position, let it indent the marked region.
Original-idea-by: Chris Allegretta <chrisa@asty.org>
And also allow it when lines are only partially indented.
This makes it possible to equalize the indentations of (accidentally)
unevenly indented lines: by first fully unindenting a group of lines,
and then reindenting them to the desired amount.
Suggested-by: Liu Hao <lh_mouse@126.com>
In this way we can delete things on the current line with ^K (after
Shift-selecting some text) without nano misdetecting a line change,
This fully fixes https://savannah.gnu.org/bugs/?52576.
Only when the cursor has moved off of the currrent line do we stop
prepending wrapped words to the next line.
This partially fixes https://savannah.gnu.org/bugs/?52576.
When the WAS_FINAL_LINE flag is relevant (when NO_NEWLINES isn't set),
the only way for 'current' to be equal to 'filebot' is when 'current_x'
is zero.
When some or all edits have been undone, and the user starts to make
new edits, the old part of the undo stack is discarded, but this does
not mean that the undo stack doesn't go back to the very beginning.
This really fixes https://savannah.gnu.org/bugs/?52504.
This also means that no question needs to be asked when exiting.
This fixes https://savannah.gnu.org/bugs/?52504.
Reported-by: Peter Passchier <peter@passchier.net>
On a freshly installed system, or for a new user, the default
XDG data directory may not exist yet. So, create it when not.
Reported-by: Brand Huntsman <alpha@qzx.com>
The softmark is now only cancelled when the cursor is moved without
Shift being held, not for things like ^L (Refresh), ^C (Position),
or ^G (Help).
This addresses https://savannah.gnu.org/bugs/?52470.
When a file is closed with the cursor on line 1, column 1, this
position is not recorded in file-positions history file -- if a
record for the file existed, it is deleted. In the latter case
the history file needs to be saved, so that other instances of
nano will know about the deletion.
This fixes https://savannah.gnu.org/bugs/?52505.
The pointer not being NULL is enough indication that the mark is set.
Also, rename the pointer from 'mark_begin' to simply 'mark', since
the former is kind of pleonastic.
Because the highlighting hinders the display of affected lines,
and, more importantly, only the highlighted part would be written
if the file was modified and the user answers yes to the "Save?"
prompt.
This fixes https://savannah.gnu.org/bugs/?52474.
Adding the obscure A_PROTECT attribute to every color prevents an
occasional miscoloring of the first few rows on the screen -- when
ncurses is version 6.0 and dates from before October 17, 2015.
Original-idea-by: Brand Huntsman <alpha@qzx.com>
Tested-by: Brand Huntsman <alpha@qzx.com>
This fixes https://savannah.gnu.org/bugs/?52182.
When there are no help lines, there won't be any call to update the
actual contents of the bottom window, so... do it immediately after
blanking the row.
This fixes https://savannah.gnu.org/bugs/?52377.
Whenever a buffer is closed, check whether the positions file on disk
was modified, and if so, reload it. Then update the position for the
closed buffer and write out the positions file to disk.
Signed-off-by: Brand Huntsman <alpha@qzx.com>
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
The function get_key_buffer() assumes waiting_mode = TRUE, but stdin
can be in non-blocking mode when a program (before nano) turned stdin
to non-blocking mode and did not change it back (possibly because it
crashed). So, explicitly set stdin to blocking mode at startup.
Signed-off-by: Lion Yang <lion@aosc.io>
When not finding a .nanorc file in the user's home directory, nano will
look for a nanorc file in $XDG_CONFIG_HOME and in the ~/.config/nano/
fallback directory. And when not finding a .nano/ subdir in the user's
home directory, nano will look for (or create) the history files in
$XDG_DATA_HOME or in the ~/.local/share/nano/ fallback directory.
This is a partial implementation of the XDG Base Directory Specification:
https://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html,
for the purpose of reducing the clutter in a user's home directory, and
to make it easier to back up just the configuration files.
Signed-off-by: Simon Ochsenreither <simon@ochsenreither.de>
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
If the user uses a single version of nano, they have no need for
--quiet. If they do sometimes use an older version and don't want
to see the warnings, they can use 2>/dev/null (they could make an
alias for that and put it before the call of nano).
Initialize 'breaking_col' to what it needs to be when the current chunk
consists of only a tab -- a tab that spreads across both start and end
of the chunk -- so that the last 'if' can move into the preceding one,
which allows the elision of 'char_len'.
In the tiny version, do_prompt() will now have an extra NULL
parameter, which will cost maybe twenty extra bytes of code.
That is acceptable when it saves thirty lines in the source.
The key sequence ^R ^X M-F <Enter> would cause nano to abort, because
it would try to add an empty string to a history list. (Furthermore,
simply don't execute an empty command, because it is pointless.)
Reported-by: Marco Diego Aurélio Mesquita <marcodiegomesquita@gmail.com>
After any replacements were made, the "not found" message is pointless
because it will be overwritten rightaway by "Replaced xx occurrences".
The message is confusing and annoying when using speech output.
This partially fixes https://savannah.gnu.org/bugs/?52282.
Reported-by: Chime Hart <chime@hubert-humphrey.com>
This fixes https://savannah.gnu.org/bugs/?52183.
Tested-by: David Lawrence Ramsey <pooka109@gmail.com>
Improved-by: David Lawrence Ramsey <pooka109@gmail.com>
Most people who make use of 'do_findprevious' and 'do_findnext'
will not make use of 'do_search' (the default binding of M-W),
so for them it is superfluous to remember the state of the
Backwards toggle in the Search menu. For the people that do
mix the usage of Alt+Up/Down with M-W, it means that M-W will
always search in the same direction as the last search. It is
a small change in behavior.
Allow the user to record and run a single macro. The default binding
for starting and stopping the recording is M-: (Alt + colon) and for
running the macro M-; (Alt + semicolon).
This fulfills https://savannah.gnu.org/bugs/?50314.
Requested-by: Peter Passchier <peter@passchier.net>
Signed-off-by: Marco Diego Aurélio Mesquita <marcodiegomesquita@gmail.com>
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
^S will be the first thing people will try for saving a file,
and ^Q is somewhat mnemonic because it is to the left of ^W:
it searches backward.
Make these keystrokes available also in the tiny version.
Before writing a file out, nano should check that the file on disk
hasn't been modified since it was read -- not only for the normal
"Write Out" action (^O), but also for "Save File" (future ^S) and
for "Save and Exit" (^X when --tempfile is used).
When writing fails and --tempfile is in effect, don't go on to prompt
for a file name; instead let the user decide what she wants to do.
This fixes https://savannah.gnu.org/bugs/?51040.
Signed-off-by: Viorel Bota <botaviorel@gmail.com>
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
When we've landed on a tab and we are moving down and the tab starts
before the current chunk, then push the index forward -- otherwise we
would not advance. When instead we're moving up and the end of the
preceding row is on the same tab as the target column AND the end of
the current row is not on that same tab, then there is some character
on this row that we can put the cursor on, so push the index forward
-- otherwise we would skip a usable row.
This fixes https://savannah.gnu.org/bugs/?52125
and fixes https://savannah.gnu.org/bugs/?52139.
In atblanks mode, if a softwrapped chunk ends with a double-width space
character (say, Unicode 003000, Ideographic Space), and that would put
half of the character past the edge of the screen, break before it.
This fixes https://savannah.gnu.org/bugs/?51671.
Simply redraw the browser screen upon every keystroke when --showcursor
is in effect, to be certain that the cursor will be placed again at the
correct spot.
This fixes https://savannah.gnu.org/bugs/?52078
and fixes https://savannah.gnu.org/bugs/?52079.
Reported-by: David Lawrence Ramsey <pooka109@gmail.com>
The basic idea is that the cursor is always off, except when it needs
to be on: when waiting for text input, and in a few other cases: when
something was searched and found in the help viewer, and in the file
browser when option -g is in effect.
This fixes https://savannah.gnu.org/bugs/?51923.
Reported-by: Mike Frysinger <vapier@gentoo.org>
When reducing the search string to at most half the screen width,
it should employ its width in columns, not its length in bytes.
Also adjust the type of a variable.
This fixes http://savannah.gnu.org/bugs/?52057.
Benno Schulenberg