aboutsummaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* remove a little more leftover code from the false startsCharles Kerr2015-05-201-8/+0
|
* copyeditingCharles Kerr2015-05-201-33/+40
|
* remove detritus from the false starts.Charles Kerr2015-05-201-362/+34
|
* fixed first test passingCharles Kerr2015-05-201-49/+156
|
* first eds test passingCharles Kerr2015-05-201-192/+76
|
* yes, generating alarms from the reference component, rather than the ↵Charles Kerr2015-05-201-49/+52
| | | | instance component, yields the right trigger timezones. Let's get to refactoring.
* what if we try calling generate_alarms() with the original component rather ↵Charles Kerr2015-05-201-4/+45
| | | | than the instance component...
* add stub for valarm repeatCharles Kerr2015-05-201-0/+7
|
* work directly with ical to generate alarm instancesCharles Kerr2015-05-191-39/+375
|
* add tracers.Charles Kerr2015-05-191-5/+23
|
* undo last Friday's timezone regression.Charles Kerr2015-05-191-1/+1
|
* in EngineEds' call to e_cal_util_generate_alarms_for_comp(), use nullptr as ↵Charles Kerr2015-05-071-1/+1
| | | | the default timezone so that we don't convert to local twice (once in _generate_alarms_for_comp, once in the DateTime ctor). Fixes the previous commit's failing test.
* resolve minor shear with trunkCharles Kerr2015-04-091-5/+4
|
* sync with trunkCharles Kerr2015-04-0910-189/+395
|\
| * in the EDS engine, give a better explanation in the comments how we handle ↵Charles Kerr2015-04-061-2/+4
| | | | | | | | alarms with no triggers, and why
| * in EngineEds, make the ECalComponentAlarmAction 'omit' array a constexpr.Charles Kerr2015-04-061-2/+2
| |
| * in SimpleAlarmQueue, add a new method 'bool already_triggered() const' to ↵Charles Kerr2015-04-061-9/+8
| | | | | | | | reduce code overlapl between find_next_alarm() and appointment_get_current_alarm()
| * in SimpleAlarmQueue, make the signature for find_next_alarm() and ↵Charles Kerr2015-04-061-29/+21
| | | | | | | | appointment_get_current_alarm() suck less.
| * in SimpleAlarmQueue, reduce a lambda capture to only the fields it needsCharles Kerr2015-04-061-1/+1
| |
| * in Actions, sync DateTime API use by calling DateTime::Local(time_t) instead ↵Charles Kerr2015-04-061-2/+2
| | | | | | | | of DateTime::DateTime(time_t)
| * in new EDS code, use timezones consistentlyCharles Kerr2015-04-061-5/+12
| |
| * in DateTime class, make it harder to accidentally mix local and nonlocal ↵Charles Kerr2015-04-061-4/+14
| | | | | | | | timezones by replacing DateTime::DateTime(time_t) with two methods, DateTime::Local(time_t) and DateTime(GTimeZone*, time_t)
| * in new code, use std::array rather than C style arraysCharles Kerr2015-04-061-2/+3
| |
| * in SimpleAlarmQueue, use references instead of copies when copies aren't ↵Charles Kerr2015-04-051-3/+3
| | | | | | | | necessary
| * make DateTime::is_same_day() fasterCharles Kerr2015-04-051-4/+6
| |
| * fix misuse of ECalComponentAlarmInstance's fields.Charles Kerr2015-04-052-5/+2
| |
| * in EdsEngine, use empty initializer lists in the new valarm codeCharles Kerr2015-04-031-10/+8
| |
| * sync with trunkCharles Kerr2015-04-031-15/+48
| |\
| | * simplify the awake display timeout codeCharles Kerr2015-03-312-12/+18
| | |
| | * the previous commit made the Awake API too complicated... simplifying.Charles Kerr2015-03-252-116/+86
| | |
| | * 60 seconds after triggering an alarm, release our keepDisplayOn requestCharles Kerr2015-03-252-50/+107
| | |
| * | add the new Alarm class as an argument to the alarm queue classCharles Kerr2015-04-034-103/+157
| | |
| * | add an Alarm class to represent ical valarm components; change the ↵Charles Kerr2015-03-315-56/+167
| |/ | | | | | | Appointment class to hold an arbitrary number of Alarms.
* | make the string test case-insensitiveCharles Kerr2015-03-241-4/+4
| |
* | add support for an x-canonical-app-url x-prop when reading vtodos/vevents ↵Charles Kerr2015-03-241-2/+22
|/ | | | from EDS
* add DateTime::end_of_month(), DateTime::end_of_day(). Add unit tests for them.Charles Kerr2015-03-162-20/+44
|
* in DateTime::is_set(), include timezone testCharles Kerr2015-03-161-2/+0
|
* in DateTime::DateTime(GTimeZone*,GDateTime*), don't allow either argument to ↵Charles Kerr2015-03-161-23/+9
| | | | be nullptr
* make utc variable name more sensible.Charles Kerr2015-03-161-2/+2
|
* fix GAction leakage found while testing new unit tests with valgrindCharles Kerr2015-03-141-0/+8
|
* in DateTime::Local() and DateTime::NowLocal(), avoid redundant construction ↵Charles Kerr2015-03-141-2/+2
| | | | of the local timezone.
* sync the rest of the code with the changes to DateTime instantiation ↵Charles Kerr2015-03-141-1/+1
| | | | mentioned two commits ago
* use the new DateTime::start_of_day() and DateTime::start_of_minute() functions.Charles Kerr2015-03-146-15/+7
|
* Add DateTime::start_of_day() to use instead of the add_hours(-hours()) ↵Charles Kerr2015-03-141-13/+64
| | | | | | | trick, which doesn't work on days when DST changes. Implementing this requires DateTime objects to keep their own GTimeZone pointer, since the one inside GDateTime is private and can't be used for DateTime::start_of_day()'s call to g_date_time_new(). As a result the public API of DateTime changes, since we'll need a GTimeZone when constructing or assigning from a GDateTime pointer.
* create a new timerfd each time we need to call timerfd_settime(). Fixes: ↵Charles Kerr2015-03-011-24/+35
|\ | | | | | | | | | | #1424966 Approved by: Ted Gould
| * create a new timerfd each time we need to call timerfd_settime().Charles Kerr2015-02-271-24/+35
| |
* | prevent deadlocking when registering G_TYPE_DBUS_PROXYCharles Kerr2015-02-281-3/+3
|/
* add a g_message() when updating due to anything that's not a periodic timerCharles Kerr2015-02-121-15/+18
|
* extract the timerfd_settime() code into its own method. Add a new call to it ↵Charles Kerr2015-02-121-20/+29
| | | | when we might have triggered via TFD_TIMER_CANCEL_ON_SET
* change the LiveClock's private impl to detect time changes from timerfd's ↵Charles Kerr2015-02-111-48/+77
| | | | TFD_TIMER_CANCEL_ON_SET, e.g. when ntp<-->manual is toggled
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-29 14:23:52 +0000