All of the callers are passing altitude in meters because that's what
pydcs uses. This still returns knots which makes it extra weird, but
that's what almost all of the callers expect.
It's probably a good idea to introduce some explicit types for the
various distance and speed units to avoid these sorts of mistakes.
The AI isn't making use of these yet, but it's not smart enough to do so
anyway.
Would benefit from an icon to differentiate it on the map.
I'm stretching the definition of "control point" quite a bit. We might
want to put a class above `ControlPoint` for `AirSpawnLocation` to
represent types of spawn locations that can't be captured and don't have
ground objectives.
Fixes https://github.com/Khopa/dcs_liberation/issues/274
"Required" SAMs (designative by redfor long range SAM launchers in the
ME) will always be spawned during campaign generation. This makes it
possible to build a semi-guaranteed IADS (the exact type of SAM is
dependent) on the choice of faction.
Requierd SAMs will consume the slots of random SAMs during generation.
Later we should differentiate between strategic SAMs like SA-10s and
tactical SAMs like SA-11s so we can fill in the medium range SAMs at
random locations among the fixed long range SAMs.
When a base is captured we clear its defenses. Those locations need to
be returned to the preset location pool so they can be used for the new
base defenses.
We currently have three methods of choosing locations for TGOs:
1. From the campaign miz
2. From the per-CP mizdata files
3. Randomly
Move the selection among these sources into a single place and use it
everywhere that we search for a TGO location.
Longer term methods 2 and 3 will be removed.
Defining a campaign using a miz file instead of as JSON has a number of
advantages:
* Much easier for players to mod their campaigns.
* Easier to see the big picture of how objective locations will be laid
out, since every control point can be seen at once.
* No need to associate objective locations to control points explicitly;
the campaign generator can claim objectives for control points based
on distance.
* Easier to create an IADS that performs well.
* Non-random campaigns are easier to make.
The downside is duplication across campaigns, and a less structured data
format for complex objects. The former is annoying if we have to fix a
bug that appears in a dozen campaigns. It's less an annoyance for
needing to start from scratch since the easiest way to create a campaign
will be to copy the "full" campaign for the given theater and prune it.
So far I've implemented control points, base defenses, and front lines.
Still need to add support for non-base defense TGOs.
This currently doesn't do anything for the `radials` property of the
`ControlPoint` because I'm not sure what those are.
Like with deleting waypoints, these will degrade the flight plan to the
2.1 behavior.
Ascend/descend points aren't in use any more, so I removed those.
In almost every case this leaves us with a flight plan we can't reason
about, so it gets degraded to `CustomFlightPlan`. The exception is when
deleting a target point when there are other target points remaining.
This probably gets people using this feature back to what they want
though, which is essentially the 2.1 behavior.
Fixes https://github.com/Khopa/dcs_liberation/issues/393
Fixes https://github.com/Khopa/dcs_liberation/issues/375
This also fixes a problem where we're spawning non-cold start planes in
an uncontrolled state. The ME won't let us do this, so we probably
shouldn't be doing that.
BAI is used for attacking ground vehicles as opposed to buildings like
strike does, and not air defenses like DEAD does. Unlike strike, BAI is
tolerant of moving targets.
Fixes https://github.com/Khopa/dcs_liberation/issues/216
This was also needed in other parts of the UI and is easier to implement
in the target class anyway.
Note that DEAD is now properly restricted to air defense targets.
Also added error boxes to the UI for when planning fails on an invalid
target.
