AI flights were only getting a single Attack Group task against multi-
group TGOs (currently only a small group of SAMs like the SA-10, I
believe), so the AI would never attack the point defense SA-15 or AAA
groups.
There is no "all but gun" RTB winchester option, so air to ground
missions with mixed weapon types will insist on using all of their
bullets after running out of missiles and bombs. Take away their bullets
so they don't strafe a Tor.
Exceptions are made for player flights and for airframes where the gun
is essential like the A-10 or warbirds.
We were setting up all the correct *target* waypoints but the AI doesn't
use the target waypoints; they use the targets property of the ingress
waypoint. This meant that the flight plan looked correct in the UI and
was correct for players but the tasks were set up incorrectly for the AI
because building TGOs are aggravatingly multiple TGOs with the same name
in the implementation.
Mission targets now enumerate their own strike targets so that this
mistake is harder to make in the future.
This won't be perfect, the AI is still not able to parallelize tasks and
since buildings aren't groups they can only attack one structure at a
time, but they'll now at least switch to the next target after hitting
the first one.
As a bonus, stop bombing the dead buildings.
Fixes https://github.com/dcs-liberation/dcs_liberation/issues/235
Fixes https://github.com/dcs-liberation/dcs_liberation/issues/244
Loadout selection no longer has two (disagreeing) implementations. What
the UI shows is now what the miz will have.
We now store the chosen layout in the Flight *always*, not just for
custom loadouts. This means that we do loadout lookups at the start of
each turn, but the data is cached in pydcs.
Era-specific loadout degradation is still done at generation (and
presentation) time. This is so that players can toggle that option and
have it affect the *current* turn, rather than the next one.
Many of the aircraft that we use for DEAD are not actually capable of
the SEAD task in DCS, so they were being loaded as some other task type,
usually one that doesn't support Attack Group, which made them lose
their waypoint actions and do nothing.
This switches them to using CAS which supports a superset of the SEAD
capable aircraft.
I've also audited the SEAD/DEAD lists. The F-117 was removed because it
is not capable of Attack Group *at all*, and all the non-SEAD aircraft
that are capable of ground attack moved from SEAD to DEAD.
Fixes https://github.com/dcs-liberation/dcs_liberation/issues/1086
This gives a clean break between the transfer request and the type of
transport allocated to make way for transports that need to switch
types (to support driving to a port, then getting on a ship, to a train,
then back on the road, etc).
https://github.com/Khopa/dcs_liberation/issues/823
The previous flight plan only makes sense if the convoy will make it a
significant distance from its starting point. At road speeds over the
typical mission duration this is not true, so we can actually plan this
as if it was a strike mission near the origin point and that's close
enough.
There's some cleanup work to do here that I've added todos for.
Fixes https://github.com/Khopa/dcs_liberation/issues/996
Still a work in progress (the missions don't actually perform their task, just orbit). Currently:
* AEW&C aircraft can be bought.
* AEW&C missions can be planned at any control point and at front lines.
* AEW&C will return after 4H or Bingo.
This makes the names of the aircraft displayed to the player in the UI more verbose and readable.
It allows allows specific countries to display an aircraft's name differently. An example of this would be the JF-17 Thunder, which is known in China as the FC-1 Fierce Dragon - this now displays correctly in the Liberation UI.
Follow up work:
* Data entry. I plan to do the air-to-air missiles in the near term. I
covered some variants of the AIM-120, AIM-7, and AIM-9 here, but there
are variants of those weapons for each mounting rack that need to be
done still, as well as all the non-US weapons.
* Arbitrary start dates.
https://github.com/Khopa/dcs_liberation/issues/490
Inexact name matching targets the first group that partially matches the
given group name. aa|71 will match aa|7.
The inexact match that was here was only needed for an early attempt to
use skynet where group names were not used consistently. That's no
longer a problem so we don't need this workaround.
Fixes https://github.com/Khopa/dcs_liberation/issues/676
Not converting all at once so I can prove the concept. After that we'll
want to cover all the cases where an int distance or speed is a part of
the save game (I've done one of them here with `Flight.alt`) so further
cleanups don't break save compat.
https://github.com/Khopa/dcs_liberation/issues/558
Moves all TOT planning into the FlightPlan to clean up specialized
behavior and improve planning characteristics.
The important part of this change is that flights are now planning to
the mission time for their flight rather than the package as a whole.
For example, a TARCAP is planned based on the time from startup to the
patrol point, a sweep is planned based on the time from startup to the
sween end point, and a strike flight is planned based on the time from
startup to the target location. TOT offsets can be handled within the
flight plan.
As another benefit of theis cleanup, flights without hold points no
longer account for the hold time in their planning, so those flights are
planned to reach their targets sooner.
As a follow up TotEstimator can be removed, but I want to keep this low
impact for 2.3.2.
Fixes https://github.com/Khopa/dcs_liberation/issues/593
This was made largely pointless in 2.2, since the AI won't plan a dozen
CAPs 300nm from the front line any more. Culling flights mostly just
confuses players and breaks the planning AI.
If we do want to limit flight counts, we need to do that by limiting
flight counts. The AI will try to put their aircraft as close to the
mission as possible, so culling will do very little to stop them from
spawning 50 flights at the front-line attached airfield.
https://github.com/Khopa/dcs_liberation/issues/578
(cherry picked from commit 0b2483ea1521af4102d98e7fda6edb0aabf81fc8)
We shouldn't be constructing these by hand, and the first argument is
not the value of the string. I'm not really sure why the current code
works at all. We probably do this in other places and should clean that
up, but for now this should fix Tauntaun.
Fixes https://github.com/Khopa/dcs_liberation/issues/528
(cherry picked from commit affb332eb9b1539bd27edc55d7d72954063d4ebf)
