When deleting an existing backup file failed, we do not want to
append to this file, but want instead to overwrite it (when the
user has put 'set allow_insecure_backup' in their nanorc file).
Also, when using O_EXCL (in the normal, secure case), O_APPEND
is pointless, because the file will be created and thus empty.
This fixes https://savannah.gnu.org/bugs/?58439.
Bug existed since version 2.2.5, commit 461519cc.
A normal user can change the group of a file (if the user is a member
of that group), but cannot change the owner of that file. So, when a
user edits a file that belongs to a different user, the call of fchown()
will fail. But there is no harm in that. Also when the user is root,
there is no harm in fchown() failing -- it will simply mean that the
backup file will remain owned by root and will not be writable by the
intended owner (when root has the normal umask of 0022).
This fixes https://savannah.gnu.org/bugs/?58383.
Bug existed since version 2.2.5, commit 86be3af7.
If unlinking would fail because the parent directory is unwritable,
then a check in has_valid_path() would have prevented do_lockfile()
from being called at all.
If in between the unlink() in delete_lockfile() and the fopen() in
write_lockfile() some other process creates a symlink in the place
of the lock file, then the fopen() could unexpectedly overwite a
root-owned file (when the user is root).
This basically reverts the previous commit, b4299f4f, but makes the
code a bit conciser.
One microsecond earlier, the lock file has been deleted (if it existed),
so, if between our unlink() and our open() some other process managed to
recreate the lock file... well, we want to delete it *again*. So, just
overwrite and truncate the lock file (if it exists).
When the lock file did NOT exist (a few microseconds earlier, when
checking in do_lockfile(), before calling write_lockfile()), then
the user expects the lock file to be written, so: just write it.
That between the check and the actual writing of the lock file there
is a small window of opportunity for other processes to write this
lock file is unfortunate, but it is not a reason to bother the user
with an error message when it happens.
One microsecond before a lock file is created an existing lock file
has been deleted, and if that deletion failed, writing the lock file
is aborted. So why should writing the lock file not be aborted when
the lock file cannot be exclusively created one microsecond later?
This makes no sense. So... always include the O_EXCL flag, also
when the INSECURE_BACKUP flag is set.
(And anyway: lock files are not an essential part of editing, they
are just a small service to the user, AND they have nothing to do
with backups, so a backup flag should not influence them.)
A file is always in either Unix or DOS or Mac format, and should
by default be saved again in that same format.
Any lone CRs or LFs after the first line should not change the
format that was deduced from that very first line.
Only when a CR is seen before any LF, and the CR is not followed
by a LF, should this CR be interpreted as a line separator. And
only then the file should be reported as being in Mac format --
as long as --noconvert is not used.
This fixes https://savannah.gnu.org/bugs/?58357.
Bug existed since at least version 2.0.6.
When the user mentions the file "zzy" on the command line,
the magic spell "Xyzzy" does not actually get typed, so the
Easter egg should then not be triggered.
When a file is saved under a different name, and as a result the
applicable syntax changes, and the old syntax had multiline regexes
and the new syntax doesn't, then the call of precalc_multicolorinfo()
in write_file() should not result in nano setting up a multicache of
zero bytes for each line in the buffer.
(Problem was found by locally letting nano crash when zero bytes are
allocated, and then happening to rename a file.py to a file.sh.)
A long option should describe what it does, not vaguely hint at it.
Also, in several places of nano's code we deal with actual temp files,
and then having a flag called TEMP_FILE that doesn't have anything to
do with temp files is somewhat confusing.
Just let all anchors disappear when doing an external spell check or
formatting or piping the buffer through an external command, because
leaving just a single anchor on the top line is useless.
This fixes https://savannah.gnu.org/bugs/?58273.
Bug existed since commit 732cf887 from one month ago. That code
causes an anchor to persist at the point where lines are removed.
Its major function is to find an applicable syntax, if there is any.
And if the syntax hasn't been used before, to prime its color pairs.
Also, reshuffle a line to be able to elide an #ifdef.
Avoid having to decrement the count in order to offset the increment
at the end of the loop.
Also, declare a variable on a separate line, and rename it.
First cutting the text to be copied (before copying it back into the
buffer) had unwanted side effects: it moved the cursor, and changed
the linestructs of the current line and the line where the mark was.
[This takes thirty more lines, but the idea is to later get rid of
the partitioning routines entirely.]
This means that in most cases mark_is_before_cursor() is called twice:
once before get_region() is called, and once by get_region() itself.
This small duplication of effort is acceptable: the affected functions
are not time critical, and it makes the code shorter.
When opening multiple files, and the first of them has a lock file,
and the user chooses to not open the corresponding file, then nano
should NOT create an empty buffer in its stead.
This fixes https://savannah.gnu.org/bugs/?57777.
Bug existed since lock files were introduced, in version 2.4.0,
specifically since commit 6948d2e7.
When at startup the user presses Cancel at the "open anyway?" prompt,
cancel the whole startup: quit. But when the user answers No, just
skip the file and continue starting up.
Don't bother looping until the 1024 bytes are read. Writing the lock
file takes just one shot too, and that is more important to get right.
Also, correct the type for the result of read(), so that -1 doesn't
get turned into a positive number, which would mean that any error
would get ignored.
It would be better if nano wrote the full filename into the lock file,
because that would be clearer when Vim displays its warning, but...
this is faster and will do for now.
(Nano should simply also store the full filename in the openfile struct,
so that get_full_path() needs to be called just once for each file.)
This addresses https://savannah.gnu.org/bugs/?57713.
The call of ferror() as parameter of a %s specifier was a mistake --
it returns a number, not a string. Avoid the problem by combining
two error checks.
The man page of fwrite() does not say anything about errno, but I guess
that the function calls write() and that the possible error numbers of
that function apply.
In theory it is now possible that fclose() fails and returns an error
that then masks an earlier error of fwrite(). But I can't be bothered:
lock files are not essential, and any errors that might occur are most
likely overlooked anyway because they are not displayed in red.
This fixes https://savannah.gnu.org/bugs/?57724.
Bug existed since lock files were introduced, in version 2.3.2,
commit bf88d27a.
All the things that we need are located within the first 68 bytes of
the .swp file. We do write byte 1007 when the file is modified, but
we never reference it. We always wrote just 1024 bytes to the .swp
file, so it makes no sense to read in any more than that.
(Vim writes 4096 bytes to a .swp at first, before adding undo stuff
four seconds later. But the 1024 bytes appear to be enough for Vim
to recognize it as a lock file.)
When asking "continue?", I always thought that answering No would mean
that nano would not continue, that it would stop, that it would abort.
But No only means that it will not open the relevant file but continues
anyway to start up. Asking "open anyway?" is more easily understood in
that way.
Unreadable or corrupt lock files are not a user error nor user intent,
so they should not keep the user from editing the corresponding file.
Also, combine some error conditions to compact the code.
This addresses https://savannah.gnu.org/bugs/?57700.
When in view mode, the file cannot be edited, so there is no need to
warn anyone (through a lock file) that the file might soon change.
This fixes https://savannah.gnu.org/bugs/?57694.
(The ideal behavior would be that nano would warn the user beforehand,
before the file that is being edited by someone else gets overwritten,
but that would require many more changes. So... just give a helpful
warning -- that's already better than it was.)
(When the new or changed name is that of an existing file AND this
file is being edited by some other editor (or nano itself), then the
behavior is rather strange: nano asks whether to continue, but the
file has already been overwritten and there is nothing the user can
do to cancel or revert things. A later commit should improve this.)
This fixes https://savannah.gnu.org/bugs/?47975,
and fixes https://savannah.gnu.org/bugs/?53883.
When we switch to another buffer, the window may have been resized
since we were last in this buffer, so make sure that 'firstcolumn'
gets a fitting value.
This fixes https://savannah.gnu.org/bugs/?56991.
Bug existed since version 2.8.0, since the softwrap overhaul.
After the previous change, all remaining calls of parse_mbchar() have
NULL as their third parameter. So, drop that parameter and remove the
chunk of code that handles it. Also rename the function, as there are
already too many functions that start with "parse".
This case is caught (by O_EXCL) later on, at the appropriate moment:
when the file is created.
Also, this removes a superfluous lstat() for each temporary file.
Weird bug. The first occurrence is probably never hit because the
existing file has been tentatively opened fourteen lines earlier,
but that looks superfluous: double work. And the second occurrence
will not be hit because the temp file is unlikely to have disappeared.
Also, don't bother assigning the stream pointer of the existing file
to variable 'f', as it gets overwritten right away by the pointer of
the temp file.
A possible change in filename won't be pushed to the screen until a
doupdate() is done, and that won't happen until nano is waiting for
a keystroke again. So... just let it be.
The command can be used to run some kind of formatter or corrector or
arranging tool on the buffer. By default the command is bound to M-F.
The formatter/corrector/arranging program must be non-interactive.
This addresses https://savannah.gnu.org/bugs/?55365,
and addresses https://savannah.gnu.org/bugs/?54651.
For the three operations (justification, spell checking, and filtering)
that need to undo or redo two suboperations (cutting and then pasting)
in one go, pass the message that describes the operation to add_undo(),
so that the undo item does not need to be modified after the call.
This looks better, but does add some twenty NULL parameters elsewhere.
The extra things that close_buffer() did are only needed and useful
when manually closing a buffer, so move them there. The other three
calls of close_buffer() only need to get rid of the current buffer
(making the preceding buffer the new current one) and nothing else.
When after reading a file we are not in curses mode, it means we have
read data from standard input, either from the keyboard or from a pipe.
In the latter case, we first need to reconnect standard input to the tty.
And in both cases, we then need to reenter curses mode before being able
to display the number of lines that were read (or an error message).
So, move the reconnecting code from scoop_stdin() to its own function,
and call this function from read_file() when needed.
This fixes https://savannah.gnu.org/bugs/?56310.
Bug existed since version 2.1.8 when reading from stdin was introduced.
The 'quiet' parameter is only used when opening a help-text tempfile
or a spell-checked tempfile, and these files necessarily exist, so
'quiet' will never be checked for them. But in the weird case that
the help-text tempfile could not be opened, it is not a problem that
"New File" gets printed to the status bar because nano will crash
shortly afterwards. And when the spell-checked tempfile could not
be opened, a "Not Found" message is appropriate, but nano will crash
in this case too.
Trying to append does not work on a fifo, and we'd like to be able to
open a fifo again. Also, the append test causes a close event for the
given file at the moment of opening it, which makes using 'inotify' for
waiting for this file to be closed useless.
Commit f8f90272 added the append test, but the original request
(https://lists.gnu.org/archive/html/info-nano/2009-05/msg00000.html
by Damien Joldersma) asked only for a warning when the user did not
have enough privileges to write to the file.
So, drop the append test and just check the access bits.
This fixes https://bugs.debian.org/583196
and fixes https://savannah.gnu.org/bugs/?29312.
The original requests (https://bugs.debian.org/551717 by Paul Wise,
and https://savannah.gnu.org/bugs/?45383 by Mike Frysinger) asked
only that specifying a directory instead of a file name should not
open a new buffer. But commit 98ffb642 excluded everything that was
not a normal file. This avoids a hang when the user accidentally
tries to open a pipe or a socket, but also prevents any user from
doing so on purpose. And opening a fifo can be useful when wanting
to handle sensitive data that shouldn't be stored on disk.
This prepares the fix for https://bugs.debian.org/583196.
Benno Schulenberg