PrusaSlicer 2.4.0-beta4

Summary

This is the fourth beta release of PrusaSlicer 2.4.0, following 2.4.0-alpha1, 2.4.0-alpha2, 2.4.0-alpha3, 2.4.0-beta1, 2.4.0-beta2 and 2.4.0-beta3. This beta mostly fixes bugs found in beta3.

To let you enjoy the beta without worries, the beta builds save their profiles into PrusaSlicer-beta directory, so you may use the beta side by side with the current release without ruining your production configuration.

Improvements with respect to PrusaSlicer 2.4.0-beta3

• Windows specific: "Dark mode" option now shows in Preferences even in standalone G-Code Viewer.
• Optimization of G-Code rendering, which improves speed issues mentioned in #7415, for example of the volumetric speed preview.
• When an existing configuration folder is found on application startup (for example configuration saved by PrusaSlicer alfa was found by PrusaSlicer beta), an extra dialog to decide whether its contents should be copied to active configuration directory was added to 2.4.0-beta2. The text in the dialog was now rephrased to be more clear. The existing configuration check is also no more performed when PrusaSlicer is started with a custom configuration folder set by the --datadir command line argument.
• Automatic random colors generated for color changes were slightly adjusted to avoid generation of extremely light or dark colors.
• We newly provide both GTK2 and GTK3 builds for aarch64 (64bit operating systems on Raspberry PI, Linux on Chromebooks). We provide both binary builds and AppImages, thanks @davidk for showing the way.
• We newly provide an AppImage for armhf (32bit operating systems on Raspberry PI) in addition to the binary .tar.bz2 distribution.

Bugs fixed

• Windows specific: Fixed resource leak that manifested by UI controls corruption and spurious crashes after application ran for a long time and parameter tabs were switched many times #7369.
• Fixed label on the new 'Upload and Simulate' button added in beta3 for Duet print host #7424.
• Windows specific, broken in beta3: With settings opened in their own non-modal window, a modal dialog opened from the settings window incorrectly opened over the main window, stealing focus from the settings window after the modal dialog was closed. This is now fixed #7419.
• Fixed a crash of the stand-alone G-code viewer when opening application preferences #7430.
• Fixed a crash in G-code preview with some OpenGL drivers #7415.
• Since PrusaSlicer 2.4.0-alpha1, first layer height could no more be specified as a percentage of a layer height. When loading an old profile, the first layer height is being properly converted to an absolute value when loaded, but the user interface still allowed the percentage to be entered. A check was newly added disallowing a percentage to be entered #7418.
• The dialog suggesting to enable 'Detect bridging perimeters' when supports are enabled kept appearing in an infinite loop under certain circumstances #7419.
• Fixed robustness issues manifested with spurious infills after re-slicing #7442.
• Fixed 'Rear' seam option, which was not always respected. This is a regression since PrusaSlicer 2.4.0-alpah1, which started to manifest after recent seam improvements.
• Fixed sliders in SLA gizmos, broken in beta2 after the 'Edit' button was added. Changing of support points diameter, hole parameters etc. was not possible for existing holes.
• Fixed reverting of a spiral vase mode setting with the "undo" button at the top of the Print settings page. Resetting configuration page to previous state using the undo button lead to a spurious dialog to change other settings in order to use spiral vase mode being shown #7146.
• OSX and Linux specific: Some of the modal dialogs with Yes and No buttons incorrectly interpreted Esc key as Yes.
• Fixed performance issues with "infill only where needed" enabled #1696 #6385.
• Windows specific: Fixed rare layout issues resulting in some of the combo boxes being rendered too small #7414, #7176.
• Fixed 'Reload from disk' which did not work for modifier volumes and which sometimes shifted the reloaded object incorrectly #7429.
• Fixed solid infill of all layers achieved by setting solid_infill_every_layers to 1. With ensure_vertical_wall_thickness disabled many layers were not filled at all, with ensure_vertical_wall_thickness enabled the layers were incorrectly filled with bridging infill #6736.
• Fixed regression in FDM supports, were the support interface at first layer overlapped with expanded support base #7460.
• Improved robustness of "don't support bridges" for bridging perimeters #7260.