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()
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.
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
Quite some time ago, many of the TBB components were deprecated in favor
of their near-equivalents in the STL or, in the case of task_scheduler_init,
were broken up and reconstituted under a less ad-hoc logic. Every time a header
file marked deprecated gets included, a rather loud warning is emitted, which
leads to a complete TBB's domination over the stderr stream during build time,
making it harder to notice _legitimate_ warnings.
Instead of merely muting the output with TBB_SUPPRESS_DEPRECATED_MESSAGES,
perform a genuine migration away from the deprecated components with the added
benefit of achieving a source compatibility with oneTBB, the successor to TBB
which has dropped the deprecated API for good.
What got replaced for what?
| Deprecated | Replacement |
| ------------------------------------- | --------------------------------------------- |
| `tbb::atomic` | `std::atomic` |
| `tbb::mutex` | `std::mutex` |
| `tbb::mutex::scoped_lock` | `std::scoped_lock<std::mutex>` |
| `tbb::mutex::scoped_lock` (empty) | `std::unique_lock<std::mutex>` (deferred) |
| `tbb::task_scheduler_init` | `tbb::global_control` |
| `tbb::this_thread` | `std::this_thread` |
Signed-off-by: Roman Beranek <roman.beranek@prusa3d.com>
structured exceptions (hard crashes, segmentation faults...),
converts them to Slic3r::HardCrash exceptions and displays the exception
using the usual PrusaSlicer way.
The SEH handler is installed on the main background slicing thread
as of now, therefore hard crashes in TBB worker threads are not handled.
from the background slicing thread, that supports cancellation.
The generic mechanism is used for generating thumbnails into G-code and
Fixes Fix deadlock when canceling the slicing while gcode is creating thumbnails #6476
Thanks @supermerill for pointing out the issue.
1) The macro-processor sanitizes the source code line for invalid UTF-8
characters. Ideally these invalid characters would be replaced with ?,
they are just dropped as of now. Better than showing an empty string
when converting to wxString though.
2) G-code export collects full error message for 1st occurence of an error
for each custom G-code block.
3) The composite error message now displays the errors collected in 2).
4) The error window is now scaled bigger and the Slicer logo is smaller
if the text is to be rendered with monospaced font, as the monospaced
text will not be word wrapped.
Implemented try/catch blocks for Slic3r exceptions and std::exceptions
with GUI error reporting. This is extremely important to report
corruption of PrusaSlicer.ini.
The following threads are named with this commit:
slic3r_main, slic3r_BackgroundSlicingProcess,
slic3r_tbbpool_xx_yy where xx is a one based index of the TTB thread
and yy is the platform thread ID.
1) All slicer's exceptions are now derived from Slic3r::Exception.
2) New exceptions are defined for slicing errors.
3) Exceptions are propagated to the Plater to show.
It remains to modify the slicing back-end to throw the new SlicingError
exceptions instead of std::runtime_error and to show the other exceptions
by a message dialog instead of a notification.
1) On Windows and Linux, the device enumeration now runs at a background
thread, while it ran on the UI thread on idle, which may have been
blocking on some rare Windows setups, see GH #3515#3733#3746#3766
2) On OSX, the device enumeration now relies on OS callback, no
polling is required.
3) Refactored for cleaner interface.
The wxString saga continues. wxWidgets 3.0 don't have the wxString::FromUTF8(const std::string&) overload, we must use the GUI::from_u8 helper
Also wxWidgets 3.0 don't allow to disable wxString->const char* conversion, so calling show_info(wxWindow*, wxString, const char*) was ambiguous
Several includes moved around
