The flag modified (io.c) keeps track of the memory state of data:
modified == 0: unchanged since load or last save
modified == 1: changed since load or last save
It is now unset in io_load_data() and io_save_cal() only.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The wins_update() call is the responsibility of the caller of
io_resolve_save_conflict().
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The thread is stopped/started in wins_prepare/unprepare_external() when hooks
are run. There is no need to do it in io_reload_data(). In fact, because of the
nested calls
notify_stop_main_thread() <--- io_reload_data()
...
notify_stop_main_thread() <--- hook/wins_prepare_external()
...
notify_start_main_thread() <--- hook/wins_unprepare_external()
...
notify_start_main_thread() <--- io_reload_data()
the thread has been running after the first hook anyway.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The introduction of hooks raised a problem with window updates. The diagnosis
in commit feb059e8 (Fix segmentation fault on reload with pre-load hook) was
right, the cure was wrong.
The problem is wins_update(), not the listbox contents. The wins_update() call
does not belong in wins_unprepare_external() (or in io_reload_data()), but at a
higher level. It should be called _after_ reload, as indeed it is in
key_generic_reload() when the listbox contents have been updated (todo as well
as appointments). The call was introduced in commit 8ae75f3 without comment.
The todo updates in io_reload_data() also belong in key_generic_reload() where
they were before commit 7f06c252.
When saving data, all panels must be updated in case a hook was executed.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Memory pointers must be initialized at the start in case memory is
freed (cleanup) before it is allocated.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
When setting start time for a new appointment, a date is disallowed
(in all other cases both time and date are still allowed).
Both date and time are displayed for the user to correct when an
appointment is changed or moved.
Built-in help in the status bar for display of input formats.
Several bug fixes that resulted in data inconsistencies (end time
before start time).
The routines use the enhanced parsing funtions to validate input:
ui_day_item_add(void)
parse_datetime()
parse_duration()
parse_datetime()
ui_day_item_repeat(void)
parse_date_duration()
parse_date()
ui_day_item_edit(void)
update_start_time()
day_edit_time()
parse_datetime()
update_duration()
parse_duration()
parse_datetime()
update_rept()
parse_date_duration()
parse_date()
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
All appointment times are checked for validity. Overflow by time
arithmetic is detected. End times are checked when appointments
are moved. Three functions are involved: parse_datetime(),
parse_duration() and parse_date_duration(); they all have a
new argument for validation purposes.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The changed handling of thread ids implies that they must be initialized
before exit of calcurse where they are used. Hence the function
vars_init() is now called irrespective of command line arguments.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
A new save thread was started every time a positive periodic save value
was input in the general options configuration menu. And only one thread
can be stopped by entering 0, also when done repeatedly.
Always stop the old thread before (possibly) starting a new.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
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>
Due to deficient validation, it is possible to get an inconsistent
database with an appointment that ends before it begins. Edit an
existing one and specify an end time by a date that precedes the start
date, or create a new one and give a date as end time that precedes the
start. Then exit calcurse; on restart it will report a data error and
exit.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
You try to enter an appointment, but enter an invalid start time (by
mistake). Calcurse rejects the input. You enter the correct start time
and calcurse asks for the description, not the end time, i.e. you get an
event.
The check for an event must only be performed on valid input.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Translations (in po-files) of texts that are used for alternative
choices (e.g. [dwmy]), may use UTF-8 encoded Unicode characters (e.g.
[éãüå]).
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Internally characters (keys) have two representations: integers and key
names. Key names are characters strings, usually the name of the
character; e.g., the character A has the representations 65 and "A", and
the tab character the representations 9 and "TAB".
The function keys_int2str() turns the integer representation of a
key/character into the key name.
For display purposes the key names are usually confined to have display
width at most three. Some curses pseudo-keys have longer key names;
e.g., the back-tab character is "KEY_BTAB". A long key name makes a
character difficult to recognize in the status bar menu.
The key name of a multibyte, UTF-8 encoded character is the conventional
Unicode name of the code point; e.g., the character ü has key name
"U+00FC" because ü is the code point 0xFC. Most of these look alike in
the status bar menu.
The patch makes the key name of a multibyte character look like that of
a singlebyte character: the character itself, i.e. the key name of the
character ü is "ü".
The main tool is implementation of a utf8_encode() routine.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Purely for readability and in preparation for the counterpart utf8_encode().
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
When a scrollbar is on display in APP or TOD windows, the right
vertical border (outside the scrollbar) is not highlighted
when the window is selected.
The scrollbar is always highlighted:
- when APP or TOD is deselected
- in configuration windows where borders otherwise are not
The patch moves the scrollbar parameters (except highlight)
from arguments of draw_scrollbar() to the function itself.
The highlight argument was 1; instead it is set higher in
the call hierarchy (wins_update_panels()) and passed on down.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Commit f8e6e0d (Fix no-colour theme, 2017-12-10) partly destroyed the
cursor in getstring() by turning it into a solid block. The fix
reintroduces wchgat() which requires a color pair argument. When no
colors are wanted, color pair 0 is used.
A similar problem exists in the layout and colour customization windows
and is fixed in the same way: move to the position and apply the reverse
video attribute.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
In keys_fill_missing() a pointer is walked through a string of
space-separated character names, but misses the string-terminating null
character.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
In colour customization, pressing cancel ('ESC' by default) will
deselect all colours and put calcurse in no-color mode. For this
to work, all colour changes must be performed with the routines
custom_apply_attr()/custom_remove_attr(). Fixed for the getstring
cursor, the scroll window border and the week number.
In addition, the week number is unconditionally coloured as the
rest of the calendar contents whether CAL is selected or not.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
When a scrollbar is on display in APP or TOD windows, the right
vertical border (outside the scrollbar) is not highlighted
when the window is selected.
The scrollbar itself is always highlighted:
- when APP or TOD is deselected
- in configuration windows where borders otherwise are not
The patch moves the scrollbar parameters from arguments of
draw_scrollbar() to the function itself.
The highlight argument to draw_scrollbar() was always 1.
Instead call circumstances are figured out and highlight set
accordingly.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
In the default colour setup (white on black), white could only
with great difficulty be used as customized foreground colour,
because the colour pair COLR_CUSTOM then was identical to
COLR_DEFAULT (default on default). This made it impossible to
distinguish the selected element in lists.
The patch turns on the video attribute bold when default is chosen
as foreground colour.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
UTF-8 encodes characters in one to four bytes (since 2003).
Because 0 is a valid code point, the decode function utf8_ord()
should return -1, not 0, on error. As a consequence utf8_width()
should return 0 for a continuation byte (as it did previously).
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Reintroduce year check for systems with a 32-bit time_t type. Remove the
lower limit (1902) for systems with a 64-bit time_t. This limits
movements in the calendar (for 32-bit systems) and in no way ensures
constistency of data.
Commit a12833e (Handle dates past January 19th, 2038, 2015-01-19)
removed the upper limit (2037) on dates but left the lower limit (1902).
It did not ensure the support of the target system.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Remove two remnants: custom_color_theme_name() and custom_set_swsiz().
Make static and rename custom_confwin_init() into confwin_init().
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Seven actions were absent from display_help(). This meant that
commands like ':help ^P' did not work. They now refer to
general.txt, but are not all described there.
For two actions, generic-copy/paste, the file name was misspelled.
This solves Github issue #44.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Previously an error from the character input routine wgetch() was
silently ignored. It should still be ignored, but must explicitly be
detected in order to be ignored. The issue came up in connection with
terminal window resize. For whatever reasons ncurses returns ERR (as
well as KEY_RESIZE) when wgetch() asks for input after a resize.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The window was not deleted if an "already in use"-key was detected,
and a new one was created as the loop was reentered.
Create/delete of the popup are moved outside the loop.
A redrawwin() call is needed to have the window displayed again.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
All keys known by ncurses can be bound. Thus the check for not
recognized keys in custom_keys_config() becomes superfluous.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
All key names for ordinary ASCII keys as well as for escape keys (pseudo
keys) are cached in a lookup-table, keynames[], at startup. Mapping
between key names (strings) and key codes (integers) in keys_str2int and
keys_int2str) is performed through this table.
The key names used are those returned by the keyname() function of
ncurses. But to accommodate some of the names to the three-letter space
available in the status menu, four ordinary keys and the most common
escape keys have calcurse abbreviated names: ESC, TAB, RET, SPC and LFT,
HOM, PgU, INS, F1, etc.
All keys known by ncurses can be bound. Thus the check for not
recognized keys in the key configuration menu becomes superfluous. The
only keys that cannot be bound, are those escape keys not known to
ncurses, i.e. not described by the terminfo database for the terminal
type in use.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Code point mapping adjusted for multibyte characters to avoid the
ncurses range 0-KEY_MAX.
This includes three fixes:
1) test sequence in keys_assign_binding(),
2) reassemble multi-byte character in keys_wgetch(),
3) check for already in use in keys_assign_binding().
Rearrangement of code. The introduction of allocated memory in
keys_int2str() has as a consequence that check for recognized ncurses
pseudo characters now are in two places: keys_str2int() and
custom_keys_config(). The latter was moved from keys_wgetch() to improve
user information.
More informative warning messages in custom_keys_config().
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The week number in the calendar panel is calculated according to
ISO 8601. Hence, Monday is the first day of the week and the
week number changes from Sunday to Monday.
However, calcurse ties the week number not to Monday, but to the
first day of the week as configured for display. Thus, when Sunday
is shown as first day of the week, the week number is correct for
Sunday, but wrong for the rest of the week (one behind).
With this patch the week number always follows the mon-sun week as
required by ISO 8601. A side effect is that when Sunday is displayed
as first day of the week, and Sunday is the selected day, the week
number displayed is invalid for the rest of the week (but changes
to the correct one when the selected day moves forward).
This raises the question whether the week numbering scheme should
follow the "first day of the week" choice and use the American week
numbering scheme instead of ISO 8601 when Sunday is the first day
of the week. But that is for the future.
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The date at the top of the appointments list may be positioned either to
the left, in the middle or to the right. Default is to the right. Can be
configured from the general options menu.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The computation never really worked before and it seems like the feature
is not very helpful, sometimes even confusing (see GitHub issue #21).
The macro ISLEAP is moved to calcurse.h.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
The reported phase was often one day early because calculations were
based on midnight. For full and new moon this is changed to noon. For
first and last quarter a direct check of the appearance of the half moon
during the selected day is performed.
Comments are relocated and expanded.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
When introducing the fmt_day_heading() function in commit d56cc7a (Make
heading in appointments panel configurable, 2017-08-19), we forgot to
initialize the dynamic string. This resulted in calling mem_free() with
a garbage pointer, potentially resulting in a segmentation fault. Fix
this by calling string_init() before using the string.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Use the locale setting to fetch the month names and abbreviated day
names.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
In the appointments panel, an appointment has a '-' to mark the time
span, and the description on the following line is slightly indented.
When the appointment is changed into a recurrent one, the '-' is changed
to a '*', making it easily distinguishable but the description also gets
a '*', thus breaking the pattern. Drop the extra '*'.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
Add a new configuration variable format.dayheading to set the format of
the date displayed at the top of the event and appointment list.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
When adding a key already in use for another action, a warning box is
displayed. The text length is limited by the window width through the
use of strncpy(). If the limit is exceeded, the string will have no null
termination, resulting in unpredictable behaviour.
A similar problem in fatalbox() is fixed as well.
Signed-off-by: Lars Henriksen <LarsHenriksen@get2net.dk>
Signed-off-by: Lukas Fleischer <lfleischer@calcurse.org>
