| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
calcurse deadlocks when
1) an upcoming appointment is on display in the notification bar,
2) an external command (like help) is started,
3) the time for the upcoming appointment arrives, and
4) the external command is exited.
The notification bar thread is stopped while the external command is
running. Upon exit from the external command, the n-bar thread is
restarted and calcurse locks.
The cause is the way in which the main notification bar thread is
stopped:
static pthread_t notify_t_main;
void notify_stop_main_thread(void)
{
if (notify_t_main) {
pthread_cancel(notify_t_main);
pthread_join(notify_t_main, NULL);
}
}
Objects of type pthread_t are opaque and should not be accessed
directly. Initially notify_t_main is an uninitialised static variable
(0), but later it has a value, which may or may not be the thread id of
the notification main thread.
Note that the thread id after exit of a thread may become the thread id
of a new thread. Thus the variable set when the thread is created, is
invalid after exit of the thread.
Specifically, the first time notify_stop_main_thread() is called (by
notify_start_main_thread() before the thread is created) is harmless
(because notify_t_main is 0). Calling notify_stop_main_thread() later
may be either
OK
because the main thread is running, or
harmless
because no thread with id notify_t_main is running: the two
functions will fail with return value ESRCH (no such process), or
fatal
because an unrelated thread with this thread id is running: it will
be cancelled, and the join may or may not succeed depending on
whether the thread is joinable or detached.
The "unrelated thread" could be the next-appointment thread,
notify_thread_app, launched by notify_check_next_app().
Always calling notify_stop_main_thread() before starting the main thread
becomes fatal when notify_check_next_app() is called shortly before
notify_start_main_thread(). This is the case in the scenario described.
The next-app-thread is then running when notify_stop_main_thread() is
called, and apparently it has the thread id of the old main thread
(confirmed by logging the return values from pthread_cancel() and
pthread_join(); the first succeeds while the second fails with EINVALID
which means that the thread is not joinable). The next-app-thread will
therefore exit without unlocking mutexes.
Ensure that notify_t_main, in case the notify main thread is not
running, has a value that it will never have when it is running. A
possibility is the thread id of the main() calcurse process (returned by
pthread_self()).
Check for this condition in notify_stop_main_thread() and set
notify_t_main when the thread is stopped.
Similar changes have been introduced for the periodic save thread and
the calendar date thread.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
| |
The configuration file (~/.calcurse/conf by default) can now be
specified with -C or --conf.
Workaround for GitHub issue #86.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
|
|
| |
The pre-load hook is often used to manipulate the data files before
loading, such as by synchronizing with a remote calendar. Make sure we
always execute the pre-load hook upon reloads, even if the data files
have not been modified.
Fixes GitHub issue #98.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
|
| |
When running calcurse in non-interactive mode, we should not expect any
user input. This is even more important in the case of iCal imports
which are used by calcurse-caldav to import events from CalDAV servers.
Partly fixes GitHub issue #73.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
After resolving a save conflict with the merge tool, we need to reload
the data files to import the result of the conflict resolution.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Turn off the local modification flag to avoid a bogus warning when
reloading the data files immediately after the merge process.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
After saving the data files, we need to recompute and store the hashes
to make sure the updated contents is reflected.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
| |
Move code to compute the hash of a data file to a separate function.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
| |
Replace the save mutex with a common mutex, which is locked whenever
read or write operations on the data files are performed. Also, since
this mutex is an implementation detail, mark the locking functions
static and remove them from the header file.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
| |
Instead of blindly reloading data in io_reload_data(), compare the
stored hashes of the data files with hashes of the current file contents
and only reload if any of the hashes differs.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
| |
Move the check to compare the stored hashes of the data files with the
current hash to a separate function. This makes the code easier to read
and reusable.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
| |
When reading the data files, compute a cryptographic hash of the file
contents and store it. When saving the files later, ensure that the hash
still matches the current file contents. If it does not, show a warning
to the user and ask whether she wants to execute the merge tool.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
| |
Using the file name extension .sav and naming the variables "backup" is
slightly misleading, since the affected files actually contain the
updated content and not some old snapshot. Use the term "new" instead.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Move the code, which runs the merge tool with current and updated data
files, to a separate function to make it reusable.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Introduce a new function keys_wait_for_any_key() and use it instead of
wgetch() whenever the return value is discarded.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
| |
Fixes GitHub issue #36.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Use mem_*() wrappers instead of directly accessing libc functions when
allocating/deallocating memory.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
|
|
| |
We need to manually force a reinitialization of the todo item list box
before reloading the items. Otherwise, the list box contains dangling
references to the linked list of todo items which has already been
cleared at this point. After the pre-load hook is called, the windows
are redrawn by wins_unprepare_external() and these invalid references
are accessed, leading to a segmentation fault.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In commit 3eae7ce (Add --list-imported command line option, 2016-01-12),
we added an option to print the hashes of imported items to stdout.
Extend this command line option such that it dumps the items using the
specified formatting strings. With the new behavior it is, for example,
easier to check items for import errors.
Also, rename the option from --list-imported to --dump-imported (it is
not part of any official release yet so we do not need to care about
backwards compatibility).
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
| |
Split io_save_{apts,todo}() into functions that write raw data to a file
and functions that write formatted items to stdout such that one can
easily extend the grep mode for format string support in a follow-up
commit.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Add a new field that indicates whether a todo item is completed or not
instead of encoding completed todo items by negative priorities.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Add a new --export-uid command line option that adds each item's hash to
the UID property when exporting.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Support "-" as file name to the -i/--import option and read from stdin
when that file is specified.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
| |
Do not show the import summary when system dialogs are disabled or when
the --quiet option is specified, even if calcurse is executed in
non-interactive mode.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
When this option is used together with -i/--import, the object
identifiers of imported objects are printed to stdout.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
| |
Implement a -q/--quiet command line option to disable system dialogs
temporarily.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
| |
Implement a new --filter-hash option to filter by object identifiers.
Each object having an identifier that has the specified pattern as a
prefix is matched. Patterns starting with an exclamation mark (!) are
interpreted as negative patterns.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
| |
Potential use case: Synchronize data with some server before loading.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds support for hooks which are executed before/after saving
calcurse data. Hooks can be placed under hooks/pre-save and
hooks/post-save in the data directory and need to be executable.
Potential use cases include:
* Automatically commit any changes to the data files using a VCS.
* Automatically sync with some sever component on data file changes.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
| |
Adds a new function io_reload_data() which can be used to reload all
appointments and TODO items.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
|
| |
Comments should only be stripped if they start at the beginning of a
line. We do not want to chop off an TODO item description or a
configuration value.
Reported-by: Håkan Jerning <jerning@home.se>
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
| |
This allows for printing a subset of the items in the data files by
using filters.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
| |
Show an error message and die if the user specified a non-existent
custom calendar file. This fixes some random hangs when calcurse is used
in non-interactive mode within scripts.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The item filters now apply to both appointments and TODO items. Also,
add a new type mask "todo" and the following new filter options:
* --filter-priority
* --filter-completed
* --filter-uncompleted
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This adds the following filter options that allow for restricting the
set of items that are read from the appointments file:
* --filter-type
* --filter-start-from
* --filter-start-to
* --filter-start-after
* --filter-start-before
* --filter-end-from
* --filter-end-to
* --filter-end-after
* --filter-end-before
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
| |
Avoid preallocating buffers on the stack, use dynamic memory allocation
instead. Also, change the semantics of new_tempfile() so that it returns
the full name of the temporary file and fix all call sites.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
| |
Use asprintf() in some cold code paths. While allocating memory on the
heap is a bit slower, using asprintf() is a bit more memory efficient
and less prone to buffer overflow errors.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
| |
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
| |
When periodic save is enabled, do not print the "The data files were
successfully saved" dialog every time io_save_cal() is called.
Reported-by: Håkan Jerning <jerning@home.se>
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
| |
If the backup file and the data file are equal, there is no need to run
the merge tool.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|
|
|
|
|
|
|
|
| |
Allow for passing an arbitrary number of arguments. This also allows us
to remove wins_launch_external2() and use wins_launch_external() at all
call sites instead.
Signed-off-by: Lukas Fleischer <calcurse@cryptocrack.de>
|