Fix LMMS plugin issues #6670
Fix LMMS plugin issues #6670
Conversation
Co-authored-by: Kevin Zander <veratil@gmail.com>
|
Works ok. Tested on Win10-64b |
|
Any updates on this? |
I'd like to complete #6680 first, then fix the merge conflicts it causes with this branch. |
|
@Testers/reviewers Here's a project file for testing #6669: And here are project files for testing #6671: NOTE: In both the master branch and this dev branch, LMMS occasionally crashes randomly during or after editing automation for the Vibed instrument. I haven't figured out how to reliably cause it yet, but I'll create a new issue for it once I look into it some more. Since it also occurs in master, it seems unrelated to anything contained in this PR and should not affect testing, but I just wanted to give a heads up in case anyone encounters it. |
|
Minor change propossed on vibed.cpp |
| m_smoothBtn.move(79, 129); | ||
| m_smoothBtn.setActiveGraphic(PLUGIN_NAME::getIconPixmap("smooth_active")); | ||
| m_smoothBtn.setInactiveGraphic(PLUGIN_NAME::getIconPixmap("smooth_inactive")); | ||
| m_smoothBtn.setChecked(false); | ||
| m_smoothBtn.setToolTip(tr("Smooth waveform")); | ||
| connect(&m_smoothBtn, SIGNAL(clicked()), this, SLOT(smoothClicked())); |
There was a problem hiding this comment.
Just a note for later enhancements: I feel we could change this to a function and just pass the variable and parameters to it.
|
Soft bump |
|
Sweet! |
This PR fixes two related bugs affecting LMMS instrument plugins, refactors the Vibed instrument, and fixes memory leaks.
Fixes #6669
Fixes #6671
Fixes #6797
1) Automated parameter loading bug (#6669, #6797)
The automation for certain parameters of LMMS plugins was failing to load correctly from the project files.
This bug affected the following parameters when they were automated:
In each of these cases, conditional parameter loading logic was used to check for the existence of a parameter with the given name in the XML data. However, only the existence of the XML attribute was checked, which is how non-automated parameters are stored. To do it correctly, they should have also checked for the existence of an XML child element, which is how automated parameters are stored. This PR fixes that oversight, which resolves the bugs.
Note: A parameter may also be stored as an attribute of a child element called "automatablemodel", but that isn't the case for any of the parameters affected by this bug.
I also took the liberty of cleaning up the style/formatting of the
loadSettingsandsaveSettingsmethods.2) Vibed data saving/loading bug (#6671)
The Vibed instrument was only saving the parameters and sample data for each of its 9 "Strings" to the project file if that particular string was active at the time the project file was saved. This led to a couple relatively serious issues including data loss.
I fixed the problem by always saving each of the 9 Vibed string parameters and their sample data regardless of whether they are currently active. I incremented the "version" attribute which Vibed XML data uses so that this change is backwards compatible with older project files.
3) Vibed refactor and improvements
The Vibed instrument's pre-C++11 codebase was in great need of an overhaul and since I was already making a number of changes to it, I decided to refactor the whole thing. Besides making the code much more readable, this refactor brings a few additional improvements/fixes such as:
4) Memory leaks
Fixes memory leaks affecting nearly all LMMS instrument plugins during heavy tempo automation. (See explanation)