Although normalized vectors in FSO are supposed to stay normalized, occasionally small floating point errors can pop up. In most cases these do not cause a problem, but in certain cases these errors can rapidly accumulate and corrupt an otherwise well-behaved matrix. This can most easily be seen with daisy-chained dock sequences or docked ships with angled dockpoints.
This PR audits the usage of `vm_vector_2_matrix_norm` to re-normalize vectors in cases where floating point errors could accumulate. There are also a few new Assertions for extra vigilance.
Follow-up to #6596. Fixes#6634.
1. Use `const char *flag_desc` because constructing a SCP_string for each flag description is not needed
2. Make `VALID_FNAME` check for nullptr
3. Modify the signature of `coalesce()` to allow capturing of mixed `const` and non-`const` arguments
4. Use `std::for_each` (via `SCP_tolower`) rather than `std::transform` to fix compilation warnings
Most places in the code that use `vm_vector_2_matrix` do so on already-normalized vectors. Changing these to `vm_vector_2_matrix_norm` avoids redundant normalization.
Several changes related to the volumetrics model fix:
1. Update the `model_load()` prototype with better parameter names, including a new `ErrorType` enum defined in pstypes.h
2. Update the `MIN`, `MAX`, and `CLAMP` macros, all of which were susceptible to performance issues because they duplicated their tokens. This could cause functions or calculations to be re-run unnecessarily. Since the codebase is now on C++17, `std::clamp()` can be used.
3. Update the rest of the codebase to fix any associated issues.
* initial version of default_settings.tbl
* Add a way to change player detail level defaults and implement a couple more options to detail_settings.tbl
* support the rest of 2d.cpp options
* Allow enforcing options to a specific value
* modernize detail defaults, levels, and values
* Fix parsing issue and remove lambda capture by reference
* fix print
* clang
* More efficient +Enforce parsing
* cleanup
* typo
* don't negate this
* error_display
* clarify language between presets, levels, and values
* add default case
* small fix
* use a map to minimize string lookups
Adds an optional argument to the `lock-perspective` sexp to allow locking of the 'free look' or 'padlock' view. This prevents the player from looking around which might not be desired in certain forced perspective situations.
Fixes a bug with new unified scanning PR (#6000) and adds a bit of cleanup, including consistent brackets, whitespace cleanup, and removing vestigial function from the original PR.
Tested and works as expected.
* Exposes the New Flight Cursor Mode to Scripting
Thanks to #6136, FSO supports switching between two methods of flying. This PR exposes those settings and values to scripters. Tested and works as expected.
* fix aiming setting
* add reset position function
* cleanup setting pitch and heading
* one more cleanup
* update to enum support
* remove older version
* one more fix for setting flight mode
* Clean unneeded properties
* properly queue-ify keys
* Refactor key checking to use functions instead of globals
* Update hooks and cache checked-against key
* Documentation and Naming
* match tilde
* only remove control not debug
This refactors the ship registry and the OSWPT struct to store data as indexes rather than pointers. This is a prerequisite for making these collections dynamic in the future. See also #5800.
Adds a `DEATH_ROLL` status to specifically cover the state after ships have been destroyed but before they exit. I've also audited all code that interacts with the ship registry to use the correct state with this new state taken into consideration.
This fixes access to exploding ships that was broken when the ship registry framework was introduced. In particular, the `when destroyed X beam-lock-all X` QoL feature now works again.
Several related fixes:
1. Copy the actual ship name, not the display name, to `killer_parent_name` for two reasons:
1a. Since `killer_parent_name` is used to look up the ship, it must be the actual ship name, not the display name.
1b. Copying the display name can cause a buffer overflow because display names can be longer than 31 characters.
2. The display name functionality is moved to the more appropriate location in `player_generate_death_message()`.
3. Remove `player_get_killer_weapon_name()` which only served as a proxy check for whether the weapon index was used or not.
4. Use `ship_registry_get()` rather than `ship_name_lookup()`.
5. Add comments to clarify the use of `killer_parent_name`.
The logic used to select builtin messages was complex and hard to follow. This commit replaces it with logic that, while not identical, is both similar and more flexible. Additionally, the old logic had a major limitation when it came to downgrading messages: If a message type was defined for *any* persona, then it couldn't be downgraded for *any* persona. The new logic removes this limitation, and makes downgrade paths easier to define as well.
Some message types had hard-coded occurrence chances. These have been removed, and the default occurrence chance for the corresponding type modified to compensate.
Add XSTR entries for "or", "None", and "Nothing".
Note that there is already an XSTR entry for "none", but it does not have the right capitalization.
Note that "None" is localized in `textify_scancode` but not `textify_scancode_universal`.
This makes "Nothing" cargo use appropriate localization.
Fixes#5227.
When iterating through `Ship_obj_list`, check the `Should_be_dead` flag before using the object. This fixes#4990 and adds defense-in-depth checks to guard against similar errors elsewhere.
I considered moving the `Ship_obj_list` removal from `obj_delete` to `ship_cleanup`, but decided it would be prudent to avoid fundamental changes to lifecycle management.
1) A comprehensive fix for the is_target_dying variable
2) Also do distance correctly for non-player target brackets by adding a check that we're dealing with the target before using the current target distance. Then also calc distance if it is not the player target.
3) Update comments and checks to be consistent
Many of these did not reflect how the variable was actually used, so update the comments.
Locale Sensitive Controls and Textfield Input
+ Add Keyboard Layout sensitive InputBoxes
+ Remove obsolete helper function
* Fig [] usage and remove old keybinding names
+ Fix Compile Warnings
+ Fix Compile Warnings 2
+ Fix Clang Warnings
+ Fix faulty style + remove lambda
+ Change reverse-lookup map to iterating function
+ Added controlconfigdefault.tbl generator
+ Fixed gcc warning
+ Fixed msvc warning
+ make debug console keyboard layout independent
+ Expand DCF Functionality for saving control presets
+ Address Issues
* Controls/update tbl (#5)
Updates controlconfigdefaults.tbl to write Primary and Secondary bindings.
* Added Scripting Hooks for controlsconfigdefault.tbl (#6)
* Handle Asserted Switch cases
* Controls is a library and should be in the libs folder
* Add Lock & EnableScripting Functionality and expose it to scripting API
* Added scripting isLocked
* Move Key Hooks to register-able callback
* Add option to remove hook
* Repurpose CCI.used for Axes
* Morph bitfield into a singleton class
* Add encoding conversion helpers
+ Add keyname conversion if needed
+ Merge refactored encoding functions
+ Change keycode encoding call
+ remove bad type modifier
+ Move Conversion functions to util/unicode
+ Change Encoding Enum to Enum Class
+ Adapt KeyName call for encoding enum class
+ Fix Warning due to enum class
+ Change validation to own function
* Make mActionToVal Assertion more helpful
* Fix typo
* Fix style
* Use constant, not magic number
* Document Ignored_keys[]
* Doxy class controls_lua_override_cache a bit better
* Put NameToVal safeties back in
* Reorder the +Disable and $Disable: options, and be exclusive
* Reorder the +Locked and $Loacked options, and be exclusive
* Revert changes in ValToJoy
* Revert errant duplicates
* Refactor printJoyJSON() to use getJsonArray for consistency
* Remove unused FS2toSDL map
* Remove ADE_FUNC(AxisActionInverted)
* Address gcc-5 issues
* Remove unused param for textify_scancode
* Declare textify_scancode_universal() in controlsconfig.h for use
* Clean up textify_scancode() and *_universal() a smidge for clarity
* Address most clang-9 issues
* Modernize NULL to nullptr
* Expand logic and document in controlsconfig::check_control() for clarity.
* Fix typo
* Remove unecassary static_cast
* Use Assertion instead of Assert within control_config_common_read_section()
* mprintf()'s need new lines!
Co-authored-by: Birk Magnussen <[email protected]>
Co-authored-by: Birk Magnussen <[email protected]>
* Fix LUA Steering Control Override
I've been using `LUA_FULL_CONTROLS` in the updated mouse-as-a-joystick script, and it works well, but there's one last small bug I'd like to fix. `LUA_FULL_CONTROLS` affects things that beyond just steering values, which has the unfortunate side-effect of also overriding non-steering controls such as cruise percent. It would be less intrusive to use the `LUA_STEERING_CONTROLS`. Unfortunately, `LUA_STEERING_CONTROLS` just sets everything to 0, which I think is a code issue bug that has been around since the lua controls were added in 2010 (https://github.com/scp-fs2open/fs2open.github.com/commit/2a69aaa23ba0def3dbfb9f566706e7130a33866f).
This PR fixes usage of `LUA_STEERING_CONTROLS`, and in testing everything works as expected (including fixing the cruise percent issue). Also, I do not think there are any scripts that use `LUA_STEERING_CONTROLS`, but if desired I can make a new version called `LUA_STEERING_CONTROLS_FIXED` or something to that effect.
* correct order of if-check
1) Don't knock the player out of the warp if the sequence has already moved to stage two. It breaks the sequence and the game will still end the mission anyway.
2) Remove the Volition code for calculating player speed. It did not actually affect the speed, but it *did* affect the calculation of warp distance which often caused the player's ship to miss the portal.
Goober5000