Increase the lines searched for PNG image #3954
Conversation
We don't have consent from the user right now and they are not encrypted. Shouldn't go to "production". BFW-3468.
…sure getting the filament out of the gear)
Dialog stores the state of underlying screen upon construction and loads it back upon destruction. Consistently. Like, always. Without this, ToolBox::DialogToolActionBox would set focused item which would prevent menu screen from serializing its focused item.
Allow them in redscreen/redscreen. It's already sent proactively, so it's safe. BFW-5207.
Show Enclosure menu only if enclosure is active Change string describing invalid temperature from "N/A" to "--"
Transfers were not cleaned up ever, if in print preview when finished, this was caused by wrongly detecting if the transfer is in progress. It did not check, if the backup file was there, do transfer, where backup file was not present (finished) were taken as in progress.
The valued that is moved from is invalidated to not reset values at the end of its lifetime. BFW-5322
This should prevent noices when calibration X axis on XL and improve Tool Offset Calibrations, but only when run as part of selftest, not as gcode (will be done in separate commit/PR) BFW-5322
Small code cleanup. BFW-5322
We shouldn't run any selftest with phasestepping enabled. To prevent running G425 with phase stepping we are disabeling phasestepping temporarily in the gcode it self instead of the selftest state machine. This improves Tool Offset Calibration results when running said calibration on already calibrated machine. The results are not perfect yet. But this change removes the major inaccuracies in the measurements. BFW-5276
We noticed shifts in measured data. Recomend reading BFW-5276 for more details, but in the end slowing down the parking speed helped a lot. BFW-5276
BFW-5335
Fill will be in the next commit BFW-5335
The problem was in write_end_item calling migrate_bank. The function is called in store init, where migrate_bank would screw things. In other cases, the free space for the end item is checked in the parent calling functions. BFW-3553
…sensor_and_user_push__ask BFW-5361
This reverts commit f605a28bea662f02ec8d5121f4f57cee8ada774b.
BFW-5297
Copy pasted from standard printing screen BFW-5371
BFW-5373
Only effects plaintext gcode. The new QOI images make it necessary to look further into the gcode to find the PNG image for display in Prusalink.
|
Thank you for the contribution. I looked into prusa3d/Prusa-Link-Web#475 and tried your fix on the file they provided. It works for the large preview, but doesn't work for the small thumbnails. The thing is, I would much rather ditch entire PNG support and let PrusaLink render the QOI files, perhaps using https://github.com/kchapelier/qoijs but that is a much larger change. I am leaving this open and will discuss this more in the team. |
|
Okay, sounds good. I hadn't realized about the small thumbnails. Where are those shown on PrusaLink? |
|
Curious if this has gone anywhere @danopernis? I know how busy the team must be so I know this isn't a top priority. But I would be interested to know if your desire of switching to QOI for everything is something that has gained any traction amongst the team? |
Only effects plaintext gcode. The new QOI images make it necessary to look further into the gcode to find the PNG image for display in Prusalink (and maybe Prusaconnect? I've not used it so not sure).
Currently, if you have a plaintext gcode file with the default gcode thumbnails (for MK4 that means: 16x16/QOI, 313x173/QOI, 440x240/QOI, 480x240/QOI, 640x480/PNG) Prusalink can't load the thumbnail, because the printer fails to send it.
This fixes prusa3d/Prusa-Link-Web#475