Move the initialization of the operating directory to after the
initialization of the screen, so that the above error can be shown.
This fixes the first part of https://savannah.gnu.org/bugs/?47798.
Between the first stat (that sets 'realexists') and the second stat
(directly after), the file might have disappeared, which would mean
that current_stat would be NULL. Prevent dereferencing this further
down.
Only when the user decides not to override an existing lockfile should
loading the corresponding file be skipped. Any failure to write the
lockfile should be ignored -- the file itself should be loaded anyway.
This fixes https://savannah.gnu.org/bugs/?47945.
Error messages about lock files should not get overwritten by purely
informational messages, only by alerting ones.
This fixes https://savannah.gnu.org/bugs/?47963.
The variable 'namecopy' has been passed to dirname(), so it is likely
to have been changed when it contains a slash. So, use a new variable
instead. Also, free the result of display_string().
This fixes https://savannah.gnu.org/bugs/?47956.
Having just opened a fresh buffer, 'openfile->next' will never be NULL,
because the list is circular.
Second, when compiled with --disable-nultibuffer, and deciding not to
override an existing lock, the 'return FALSE' should *not* be skipped,
because otherwise the named file will be opened after all.
This fixes an unreported bug.
Either the selected file was changed, or a file or directory was chosen
(and the Enter key virtually pressed), or a shortcut was executed. So,
just go and read the next key.
This fixes https://savannah.gnu.org/bugs/?47944.
If during startup there are multiple error messages, currently only the
last one remains and can be read. To improve on that, introduce a short
pause between error messages -- even if it's not enough to read them all,
at least the user will be aware that there are multiple ones.
This also causes a few error messages to beep that currently don't beep,
such as when a file is unwritable.
When doing a PageUp or PageDown in the browser, don't move the highlight
to the first line in the same column, but keep it in the same relative
position of the screen. If we're already on the first or last page,
move the highlight to the first or last line, but keep it in the same
column. If we're already on the first or last line, only then move it
to the first or last entry.
CONST_UPDATE is only relevant when in the main menu, not when in the
browser. So, check for that condition, instead of saving, changing,
and then restoring the setting.
There is no need to iterate through all the file names, as we already
know the length of the longest. From this it's easy to calculate how
many names fit on a line.
Also rename some things to be more fitting. And make sure that something
will be selected even when many files, including the selected one, have
disappeared.
When refreshing the screen (and thus the file list), use the prev_dir
mechanism to reselect the file that was selected before the refresh,
to prevent the selection from changing when files were added or deleted.
Also, when the selected name has disappeared, move the highlight one step
back, so that it is obvious that the selection has changed. (Decrementing
'selected' will never make it negative, because selected == 0 means the
'..' entry, and every directory has a '..' entry, so it will be found.)
This fixes https://savannah.gnu.org/bugs/?47812.
Signed-off-by: Rishabh Dave <rishabhddave@gmail.com
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
When during searching we look at the keyboard, then don't look at
just the first keystroke, but check if there are more, and if so,
check each one until we find a Cancel -- or until all keystrokes
have been consumed and we can continue.
This fixes the first half of https://savannah.gnu.org/bugs/?47438.
Cancelling a search should restore waiting mode for the keyboard,
just like at the other exit points of the search routine.
And when the input routine has read in multiple keystrokes, it
should not blindly switch back to waiting mode, but only when
that mode was on before.
This fixes the second half of https://savannah.gnu.org/bugs/?47438.
When the terminal is very narrow, there is little point in showing only
part of the version string -- and chewing off one or two digits from the
version number might even give someone a wrong idea. The user is better
served with always showing the full filename, as long as it fits in the
available screen width.
This fixes https://savannah.gnu.org/bugs/?47703.
When fixing a misspelling and not answering 'All' (but 'Yes' or 'No'),
the chance is quite substantial that one will type y or n at the end of
the next word. Diminish this chance by showing for a moment a message
on the statusbar.
Partitioning the file makes the undo system lose track, so that undoing
things has a good chance of losing data. Instead, just make sure that
the region is marked "backwards", with current at the top and the mark
at the bottom, and then let the replace loop take care of not going
outside of the marked region.
This also has the effect that if the marked region fits entirely on the
screen, or all the misspellings are onscreen, then the screen will not
be scrolled at all. Which makes for a smooth experience.
This fixes https://savannah.gnu.org/bugs/?47836,
and fixes https://savannah.gnu.org/bugs/?45573.
Since nano-2.4.1, reading in or pasting a large piece of text would put
the cursor on the bottom line, leaving only one line of the non-read or
non-pasted text visible. This is different from the centering behavior
of Pico, and somewhat disorienting, as you can't see "where you are" any
more in relation to the file as it was.
So now center the cursor whenever the read or pasted text is larger than
the screen, but don't center it when the text fits entirely on the screen.
(The latter avoids the effect of the screen jumping unnecessarily when
inserting just a few lines while the cursor is near the bottom.)
To achieve this behavior: default to focusing, and temporarily set it to
FALSE when the focusing effect is unwanted.
This fixes https://savannah.gnu.org/bugs/?47841.
For some reason the keyboard code seems to assume that all integers
waiting in the keybuffer form a single sequence. But this isn't the
case: when there is an escape (0x1b) among them, then that is where
a new sequence starts. So, prevent the input code from considering
an escape after a non-escape as part of the current sequence.
This fixes https://savannah.gnu.org/bugs/?47839.
When starting to replace misspelled words, pass the actual position
of the cursor instead of the position of the first found occurrence,
so that the cursor position will be updated for changed lengths.
This fixes https://savannah.gnu.org/bugs/?47834.
When verifying that a match is a separate word (during spell checking),
instead of first copying out the word, then passing the word, and then
measuring its length, just pass its length straigtaway.
When we're spell checking, we don't need a special mechanism to detect
we have come full circle: reaching the end-of-buffer means we're done.
So don't bother to reset came_full_circle when we're spell checking
(when begin == NULL) but simply ignore its value.
Add a global variable, 'present_path', so that 'cwd_tab_completion()'
knows where the user is in the browser, so that it can try completions
against names in that directory instead of always against names in the
current working directory (where nano was invoked).
This fixes https://savannah.gnu.org/bugs/?47234.
Signed-off-by: Rishabh Dave <rishabhddave@gmail.com>
Signed-off-by: Benno Schulenberg <bensberg@justemail.net>
The internal spell checker starts searching/replacing always at the top
of the buffer, so reaching the end of the buffer means we're done with
the current search/replace. This prevents the "Search Wrapped" message
from flashing over the statusbar at the end of a spelling replacement.
Commit 8704dde mistakenly removed this part of code -- it is not dead,
it is just that it will only fire when the user answered No at some of
the replacement prompts. So... when we've rereached the starting line,
a found occurrence is invalid when it is beyond the starting x (either
after or before it, dependending on the direction of search).
This fixes https://savannah.gnu.org/bugs/?47816.
Included files are meant to contain only syntax definitions; they may
not set or unset options, nor rebind keys, nor include other files.
This fixes https://savannah.gnu.org/bugs/?47391.
For a little contrast with the function edit_refresh() -- it's
annoying that when you search for the latter you get to see all
the settings of the flag too.
The function edit_update() is called by edit_refresh() itself, so it is
silly that the first sets 'edit_refresh_needed' to TRUE. This setting
is needed only in a few cases -- in the others it's not needed because
the screen does not need to be refreshed (it was just about positioning
the cursor), or 'edit_refresh_needed' has already been set by a call to
goto_line_posx(). So, just set the flag in the five places that need it
and spare the other four calls.
Since commit 41ed690, cancelling a prompt after tabbing would sometimes
leave the list of file names on the screen. When testing this first,
it worked fine -- I was fooled again by 'edit_refresh_needed' already
being TRUE when nano has just started up and sits waiting for the first
key stroke. I have to hunt this down and kill it.
Since commit dac0aa1, nano would leave edittop at current after reading
data from standard input, meaning that everything that had been read in
was invisible, "scrolled off" the top of the screen. Correct this by
explicitly setting edittop.
The above does not happen for ^R (^X) because it has a special mechanism
to save and restore edittop. Nor does it happen for ^R ^X M-F because it
sets current to the top of the file and refreshing the screen will make
edittop follow, or -- in case just one unterminated line was read in --
edittop will already be equal to current.
A normal lock file is apparently 1024 bytes in size, so the second
attempt at reading bytes from the file would try to read 8192 more
bytes into a buffer that has room for only 7168 left. According to
valgrind, the read() function doesn't like that -- and true: if for
some reason the lock file had suddenly expanded, the buffer would
overflow.
This fixes https://savannah.gnu.org/bugs/?47156.
Commit 36ec76a made the wrong change: after a tab that did not list any
file names on the screen, a refresh /is/ needed, because a previous tab
might have listed things on the screen. But at the end of the prompt,
it is not necessary to refresh the edit window if things were listed,
because the window will be refreshed anyway after reading in a file.
Use 'slash' to point at a possible slash, use 'filename' just to
point at the real file name, and use 'wasdirname' just to point at
the dir's name before expanding it in order to be able to free it.
Also, remove two superfluous asserts: 'dirname' cannot be NULL
because it has just been mallocstrcpy'd, and checking 'num_matches'
is pointless as it would crash on the next statement anyway.
This is a remnant from 2001, when things were different. Now, there
is no need to refresh the edit window when tabbing produced no list.
When it did produce a list, it is cleared off later.
If for some reason opening the spell-checked or formatted file fails,
don't throw away the current contents of the buffer.
(It should also give proper feedback about the failure, but we'll leave
that for some other time.)
Also, store the input character earlier, so we don't have to use len - 1.
Furthermore, len increments in steps of 1, so it cannot pass the value of
bufx unnoticed, so use a comparison for equality.
Most of the time NO_CONVERT will not be set, the number of lines will
not be zero, and the format of the file will be zero. Rearrange the
conditions so that they will evaluate as FALSE as soon as possible.
Index i follows almost synchronously the value of len. Since we're
adding characters to the intermediate buffer always only at the end,
just use len as the index.
Until now (when not leaving files unconverted), nano would fumble and
drop the final carriage return of a Mac file, and would thus treat the
last line of such a file as an unterminated line and prepend it to the
current line of the buffer. Correct that, and delete the dead piece
of code that was meant to do this.
This fixes https://savannah.gnu.org/bugs/?47716.
When we don't set edittop in read_line(), we don't need to readjust it in
read_file(), because in that particular case it will still be pointing at
current. And since fileptr is a new, freshly created line, it can never
be equal to filebot, so there is no point in comparing them.
If more than one line was inserted at the beginning of the file, leave it
up to the screen handling to set edittop to what it should be.
Move the setting of fileage a bit down, to its sister setting: the line
at current gets "connected" either to the top-of-file pointer or to the
last line of the inserted file.
Benno Schulenberg