flip direction of the infill layer with each interface layer.
Unfortunately the flipping of support interface directions may not work
reliably due to base support layer heights growing at different rate
from the interface layers.
for offset2() with clear meaning.
New ClipperUtils functions: expand(), shrink() as an alternative
for offset() with clear meaning.
All offset values for the new functions are positive.
Various offsetting ClipperUtils (offset, offset2, offset2_ex) working
over Polygons were marked as unsafe, sometimes producing invalid output
if called for more than one polygon. These functions were reworked
to offset polygons one by one. The new functions working over Polygons
shall work the same way as the old safe ones working over ExPolygons,
but working with Polygons shall be computationally more efficient.
Improvements in FDM support generator:
1) For both grid and snug supports: Don't filter out supports for which
the contacts are completely reduced by support / object XY separation.
2) Rounding / merging of supports using the closing radius parameter is
now smoother, it does not produce sharp corners.
3) Snug supports: When calculating support interfaces, expand the projected
support contact areas to produce wider, printable and more stable interfaces.
4) Don't reduce support interfaces for snug supports for steep overhangs,
that would normally not need them. Snug supports often produce very
narrow support interface regions and turning them off makes the support
interfaces disappear.
Fixed crash with zero support base spacing.
The bug fix was not correct and it disabled the new "zig-zag" sparse
infill generator, leading to GH issue #7014
Somehow improved missing interface layers for snug supports
by propagating full overhangs when generating interface layers.
Fixed generation of soluble interfaces for support enforcers,
where base support was used for steeper overhangs.
Disabled filtering out thin regions from the lower layer, that will
not be covered by perimeters, thus they are not supporting the current layer.
However this may lead to a situation where regions at the current layer
that are narrow thus not extrudable will generate unnecessary supports.
For example, see GH issue #3094
* Fixed update of a warning icon for multi-part object, when some part with errors was deleted
* Fixed a crash when selecting non-Object and non-Volume item (in get_mesh_errors_info() was missed check of the obj_idx)
+ Follow-up 1cc7b4ba97 - some code refactoring:
* fixed RepairedMeshErrors.merge()
* get_mesh_errors_count() renamed to get_repaired_errors_count()
with their original names. Using weird shorthands makes the code
unreadable for anyone but the original author.
template<class T> using shptr = std::shared_ptr<T>;
template<class T> using uqptr = std::unique_ptr<T>;
template<class T> using wkptr = std::weak_ptr<T>;
There was a bug for more than 4 years, which likely triggers now more
often with the introduction of raft_contact_distance parameter,
which is usually significantly smaller than support_material_contact_distance.
There were no support towers built at the raft for contact layers
below the print_z of the first object layer.
1) Added parsing of 3MF PrusaSlicer generator semantic version
2) For 3MFs generated by >= "2.4.0-alpha1" and < "2.4.0-alpha3",
remove unreferenced vertices on loading. This should shrink the incorrectly
generated 3MFs back.
3) Added check for an empty mesh, which could have crashed PrusaSlicer on 3MF parsing.
This is a regression wrt. PrusaSlicer 2.4.0-alpha1 due to the G-code
processing optimization and parallelization.
Related to GCode Viewer changes files modified date in windows #5079
Send a warning notification if this happens. Also,if there is no current SLA profile initialized, an error message will tell the user to switch to an SLA profile before importing.
fixes#6915
Volume transformations were ignored in SLA mode. This did not matter for plain STLs
and PS own 3MF, because in those cases, the volume trafo was identity. Importing
a 3rd party 3MF leads to issues with support/holes placement and generation.
Fixes#6100 and #6744.
* Check Unsaved changes (partially related to #5903)
+ Allow create new project when Plater is empty, but some of presets are modified (related to #5903)
+ When creating new project allow Keep or Discard modification from previous project
+ Added check of changes:
* before any load project (including DnD and "Load From Recent Projects")
* before preset updater
* when configuration is changing from the ConfigWizard
+ Dialog caption is added for each check
+ Create/Destroy ConfigWizard every time when it's called
* Check Unsaved changes: Next Improvements
+ For dialog "Save project changes" added a reason of saving and name of the current project (or "Untitled")
+ UnsavedChangesDialog: Headers are extended to better explain the reason
+ Preferences: Fixed tooltiops for "Always ask for unsaved changes when..."
+ Suppress "Remember my choice" checkbox for actions which are not frequently used
* Fixed behavior of the application when try to save changed project but "Cancel" button is selected in "Save file as..." dialog
* Check unsaved changes: Improvements for Config Wizard - Check all cases when presets should be updated
+ Fixed info line for Materials pages. Text of the info relates to the printer technology now
* Improved suggested name for a project when Application is closing
* Fixed Linux/OSX build warnings
Zero support base spacing is newly allowed, switching from
the "support base" infill pattern to rectilinar infill pattern.
Why someone would want to use a solid infill for support base eludes me,
but it is simpler to support it instead of working out some rules on
minimum support base density.
Fixes Support patern spacing set to zero causing crash #6989
+ Show 2 groups of errors: "Auto-repared" and "Remaning".
+ Use different icons for volumes with/without remaining errors
+ Fixed update of the ObjectList warnings icons after switch the color mode of a PrusaSlicer
+ Some code refactoring: Use same code for tooltips for "exclamation"-icon from an ObjectList and from a Sidebar->ObjectInfo
so that a mesh with incorrectly oriented faces will not be considered
a manifold.
Added assert for 3D convex hulls constructed by QHull. They shall be
manifold (however sometimes they are not).
TriangleMesh newly only holds indexed_triangle_set and
TriangleMeshStats. TriangleMeshStats contains an excerpt of stl_stats.
TriangleMeshStats are updated when initializing with indexed_triangle_set.
Admesh triangle mesh fixing is newly only used when loading an STL.
AMF / 3MF / OBJ file formats are already indexed triangle sets, thus
they are no more converted to admesh stl_file format, nor fixed
through admesh repair machinery. When importing AMF / 3MF / OBJ files,
volume is calculated and if negative, all faces are flipped. Also
a bounding box and number of open edges is calculated.
Implemented its_number_of_patches(), its_num_open_edges()
Optimized its_split(), its_is_splittable() using a visitor pattern.
Reworked QHull integration into TriangleMesh:
1) Face normals were not right.
2) Indexed triangle set is newly emitted instead of duplicating
vertices for each face.
Fixed cut_mesh(): Orient the triangulated faces correctly.
while the layer height profile vector may not contain all print z
values, the only real case with two eqiuvalent layer height profiles is
when one was created as a copy of another, so the vectors should in fact be equal.
stl_stats are newly only accessed by TriangleMesh::stats(),
most of the direct access to TriangleMesh::stl is gone with the exception
of parsing input files (3MF, AMF, obj).
Sometimes Clipper produces a polyline with more than 2 points when
clipping a line with a polygon or a set of polygons. We hope the intermediate
points are collinear with the line, so we may just ignore them.
Exporting G-code on a worker thread did not work correctly as the worker
threads were using user's locale, not "C" locale.
The "C" locale is newly enforced to TBB worker threads by
name_tbb_thread_pool_threads_set_locale()
WIP to G-code export parallelization through pipelining:
Decoupled CoolingBuffer from GCode / GCodeWriter, ready to be
pipelined on a different thread.
WIP to G-code export parallelization through pipelining:
GCodeProcessor is called during the G-code export,
the G-code is no more reopened and re-read, but it is pipelined
from the G-code generator.
GCodeViewer no more parses G-code just to extract line end positions.
Removed start_mapping_gcode_window(), void stop_mapping_gcode_window(),
they are no more needed.
The old version of GCC and Clang support only integers to be passed to std::to_chars and std::from_chars. macOS older version of Clang doesn't support std::from_chars at all. So for Linux and macOS, it was replaced std::from_chars with strtod and temporarily was replace std::to_chars with snprintf.
This is a regression to a late PrusaSlicer 2.4.0-alpha0 change
8dfc0422a8
Faster and hopefully more reliable projection of paint-on support
blockers and enforcers on a sliced mesh.
Previous d89f01c717 did not fix it.
This is a regression to a late PrusaSlicer 2.4.0-alpha0 change
8dfc0422a8
Faster and hopefully more reliable projection of paint-on support
blockers and enforcers on a sliced mesh.
../src/libslic3r/QuadricEdgeCollapse.cpp:628:21: warning: comparison of integer expressions of different signedness: 'const int' and 'uint32_t' {aka 'unsigned int'} [-Wsign-compare]
../src/libslic3r/QuadricEdgeCollapse.cpp:631:21: warning: comparison of integer expressions of different signedness: 'const int' and 'uint32_t' {aka 'unsigned int'} [-Wsign-compare]
../src/libslic3r/QuadricEdgeCollapse.cpp:638:48: warning: comparison of integer expressions of different signedness: 'const int' and 'uint32_t' {aka 'unsigned int'} [-Wsign-compare]
../src/libslic3r/QuadricEdgeCollapse.cpp:643:25: warning: comparison of integer expressions of different signedness: 'const int' and 'uint32_t' {aka 'unsigned int'} [-Wsign-compare]
../src/libslic3r/QuadricEdgeCollapse.cpp:647:25: warning: comparison of integer expressions of different signedness: 'const int' and 'uint32_t' {aka 'unsigned int'} [-Wsign-compare]
1) Activate installed filament or SLA material profile after update_compatible(),
so that the compatiblity and visibility flags of presets are updated.
2) Only activate the first newly installed filament / SLA material profile
if the active printer did not change. This also means that if no filament
profile was active before Wizard was open or it became incompatible with
the newly installed Printer profile, the default filament profile assigned
to the activated Printer is activated preferably, which may or may not
be one of the newly installed filament profiles.
object layer over raft interface:
"first_layer_speed_over_raft", "first_layer_acceleration_over_raft".
Fixes I have a question about the speed of the first layer after the raft. #6623
Fixes Layer After Raft Is Not Considered First Layer! #6166
1) Changed the name of the variable "brim_offset" to "brim_separation"
for clarity.
2) Added legacy conversion after loading an old 3MF that does not define
then new "brim_separation" variable: The "brim_separation" is being
filled in with the "elefant_foot_compensation" value to produce
equal brim separation to the old PrusaSlicer that saved that 3MF file.
reset timestamp to 1. This led to a bug where e.g. deleting painted facets
through the respective item in object list followed by possible other actions
and undo restored the painted facets from the time when the project was loaded.
I'm not sure if there was any other situation where this problem manifested.
The triangle-ray intersection function used a hard coded epsilon,
which did not work for triangle meshes, that were either too small
or too large. Newly the epsilon may be provided to the AABBTreeIndirect
search functions externally and IndexedMesh calculates a suitable
epsilon on demand from an average triangle mesh edge length.
6b03b66167
Renamed the new "printhost_ignore_check" option to "printhost_ssl_ignore_revoke"
Improved the Physical Printers dialog in regard to the new option checkbox
(added tooltip to the checkbox, moved it to the end of options).
Disabled the host_xxx options at the command line interface, they no
more work after these options were separated to Physical Printers profiles.
Little refactoring of Http.cpp/hpp, OctoPrint.cpp/hpp
Private local variables prefixed with m_, some renaming for clarity.
list of modified options.
Optimization of DynamicConfig::equals(), ::diff(), ::equal()
to iterate over the two compared std::map trees instead of
first generating a list of keys and then searching for each key
in the respective map.
Optimization of PresetCollection::current_is_dirty() and ::saved_is_dirty()
to call DynamicConfig::equals() instead of ::diff().
of a printer from MMU to non-MMU (when number of extruders changes).
Also slightly optimized in case a MMU-painted object is sliced in
single extruder mode.
PresetComboBoxes: Auto selection of printer/material preset from the new added presets in ConfigWizard
+ Check unsaved changes if ConfigWizard is running from the PesetComboBoxes
1) Slic3r::RuntimeError was replaced with ConfigurationError,
all exceptions thrown by the configuration layer are derived
from ConfigurationError.
2) When parsing configuration files, ConfigurationError is catched and
rethrown extended with the file name being parsed.
1) Starting with this commit, configuration block exported into G-code
is delimited by "; prusaslicer_config = begin" and "; prusaslicer_config = end".
These delimiters look like any other key / value configuration pairs
on purpose to be compatible with older PrusaSlicer config parsing from G-code.
2) Config parser from G-code newly searches for "; generated by ..."
comment over the complete G-code, thus it is compatible with various
post processing scripts extending the G-code at the start.
3) Config parser from G-code parses PrusaSlicer version from
the "; generated by PrusaSlicer ...." header and if the G-code was
generated by PrusaSlicer 2.4.0-alpha0 and newer, it expects that
the G-code already contains the "; prusaslicer_config = begin / end"
tags and it relies on these tags to extract configuration.
4) A new simple and robust parser was written for reading project configuration
from 3MF / AMF, while a heuristic parser to read config from G-code located
at the end of the G-code file was used before.
- The optional output_name file might contain the .pp suffix.
- In case the file contains just filename, prepend it with the
output dir. We don't want to save to current workdir.
1) New environment variable SLIC3R_PP_HOST contains one of
"File", "PrusaLink", "Repetier", "SL1Host", "OctoPrint", "FlashAir", "Duet", "AstroBox" ...
2) New environment variable SLIC3R_PP_OUTPUT_NAME contains the name
of the G-code file including path (for SLIC3R_PP_HOST == "File")
or a name of the file after upload to the host (PrusaLink, Octoprint ...)
3) The post-processing script may suggest a new output file name
(likely based on SLIC3R_PP_OUTPUT_NAME) by saving it as a single line
into a new "output name" temp file. The "output name" file name is
created by suffixing the input G-code file name with ".output_name".
Please note that the G-code viewer visualizes G-code before post-processing.
Fixes Broken PostProcessing when script changes out-filename #6042
Make is_converted_from_meters / is_converted_from_inches exclusive-or.
Maybe it would be better to make a single enum from the two booleans,
if they are exclusive-or?
It is now possible to use e.g. --ensure-on-bed=0 for bools (meaning the same as --no-ensure-on-bed).
Using --no- prefix on non-boolean is an error (--no-ensure-on-bed=1)
Providing a value for --no- prefixed bool is an error (--no-loglevel 5)
'dont-ensure-on-bed' (which allows to override). This was the original
behaviour in Slic3r and Sli3rPE, probably broken long ago when CLI
was ported from Perl.
Also, --scale-to-fit should now work again (#5772)
Filtering of unprintable regions in multi-material segmentation depends on if gap-fill is enabled or not. So sliced object is invalidated when gap-fill was enabled/disabled by option "gap_fill_enabled" or by changing "gap_fill_speed" to force recomputation of the multi-material segmentation.
For Apple's on Arm CPU computed triangle normals inside fragment shader using dFdx and dFdy has the opposite direction. Because of this, objects had darker colors inside the multi-material gizmo.
Based on https://stackoverflow.com/a/66206648, the similar behavior was also spotted on some other devices with Arm CPU.
1) Polished up wording of the error messages.
2) Made some messages in the SysInfo dialog localized.
3) Renamed LibraryCheck.cpp/hpp to BlacklistedLibraryCheck.cpp/hpp
4) CPPized the BlacklistedLibraryCheck WIN32 C code.
It was reworked graph generation for multi-material segmentation. Now only oriented arcs in one direction are added to the graph for input polygons. This direction matches the direction of the lines in the input polygons.
To the "physical_printer" added "preset_name" option as a duplicate of the "preset_names" but like one string.
Note: It will works just for printer preset names without ";" (see #5503)
