Skip to content

Use final value instead of value to reflect automation in get_parameters - #442

Closed
Claers wants to merge 1 commit into
utopia-rise:masterfrom
Claers:fix/get_final_value_get_parameter
Closed

Claers wants to merge 1 commit into
utopia-rise:masterfrom
Claers:fix/get_final_value_get_parameter

Conversation

@Claers

@Claers Claers commented Aug 12, 2026

Copy link
Copy Markdown
Contributor

When updating a parameter in FMOD (for example to have a timing for the engine to prepare an event) the actual code is using value but in FMOD Engine value is the value that is set by the user and does not reflect actual value changes made by FMOD itself.

This pull request change this by using final_value instead of value. This should not break things bc it's a voluntary action to change a value into FMOD and want to gather it into the engine.

Shaun-Fong added a commit to Shaun-Fong/fmod-gdextension that referenced this pull request Oct 19, 2026
Fixes taken from the open pull requests of utopia-rise/fmod-gdextension:

* utopia-rise#443: keep the UTF-8 converted plugin path alive while FMOD loads it
* utopia-rise#442: return the final parameter value from get_parameter_by_* and fix
  the error messages of the global parameter getters
* utopia-rise#441: do not restart autoplay events when the editor sets the event
* utopia-rise#429: do not dereference a null event description and do not build a
  garbage path when getPath returns FMOD_ERR_EVENT_NOTFOUND
* utopia-rise#417: look the plugins settings up with ResourceLoader so exported
  projects find them
* utopia-rise#434: link with a weak load_all_fmod_plugins on iOS so a project
  without static plugins still links (the rest of utopia-rise#434 is an iOS
  simulator build overhaul and is out of scope here)
* utopia-rise#395: validate the event description before reading parameters

Bugs found while auditing the code base:

* FmodServer::_set_listener_attributes returned instead of continuing on
  a listener outside the tree, so the remaining listeners were skipped
* FmodServer::shutdown released the FMOD system before the event
  instances and deleted the cache while banks could still be referenced,
  leaving dangling handles and crashing on late Ref releases;
  FmodBank::~FmodBank now also checks the server singleton
* load_bank, unload_bank and banks_still_loading dereferenced the cache
  without a null check, and unload_bank refused to unload banks that are
  packed in the PCK because of a FileAccess::file_exists check
* get_global_parameter_desc_list truncated its result to 256 entries and
  the performance counters overflowed after 2 GiB of reads
* The asynchronous file reader waited on a condition variable without a
  predicate and could miss a notification, which blocked the read thread
  forever; the thread now drains pending requests when it stops
* godotFileOpen dereferenced the Ref returned by FileAccess::open even
  when the open failed, and godotFileClose dereferenced a null handle
* The event callbacks dereferenced the server singleton, the event user
  data and the programmer sound without null checks
* FmodEventDescription::get_instance_list built a Ref from an
  uninitialized pointer, and the parameter label getters returned an
  uninitialized buffer when the FMOD call failed
* string_to_fmod_guid left the GUID uninitialized when the string did not
  parse, and FMODCLASSWITHPATH read the path of a failed getPath
* FmodEvent::set_node_attributes dereferenced a null node and
  FmodSound::is_valid tested an uninitialized bool on a null handle
* The logging settings enum hint was off by one, so DEBUG_INHERIT was
  missing from the inspector and DEBUG_VERBOSE could not be selected
* D_METHOD argument name typos in FmodEventEmitter and
  FmodEventDescription

Release workflow: replace the archived node12 release actions
(actions/create-release@v1 and actions/upload-release-asset@v1) with
softprops/action-gh-release@v2, use checkout@v4 and grant contents: write
so the tag build can publish its assets.
@CedNaru

CedNaru commented Nov 3, 2026

Copy link
Copy Markdown
Member

Superseded by PR #452

@CedNaru CedNaru closed this Nov 3, 2026

This branch had an error being deployed

1 failed deployment
external-pr — 1b0023aa Deployed Aug 12, 2026 by Claers via gate #485
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants