# quick fix is to rerun the last link step with added -latomic # /usr/include/c++/8/atomic:239: undefined reference to `_atomic_store_8 # if you compile using gcc g++ then the build stop with the following exception: DSKIA_LIBRARY=$HOME/deps/skia/out/Release-ARM32/libskia.a \ DSKIA_LIBRARY_DIR=$HOME/deps/skia/out/Release-ARM32 \ Git clone -depth 1 -branch beta -single-branch -recursive #clone the asesprite beta branch, without history #create the deps folder with precompiled skia for 32bit arm Sudo apt-get install -y clang cmake ninja-build libx11-dev libxcursor-dev libxi-dev libgl1-mesa-dev libfontconfig1-dev libbb2-dev # about 2.5Gb of disk space will be consumed during the build # 2Gb of ram is required to complete the build # lines without # are to be copied and run on the raspberry pi command line My precompile version of skia for Raspberry Pi ARM Linux 32bit can be found at: Ĭompile instructions # lines starting with # are comment Both HDMI out and the official 7" TFT work. Testing and running aseprite 1.3.x beta on various Raspberry Pi models. Gn, ninja and skia that would allow you to compile and perform the final linking of the aseprite binary in about 30 minutes on a Raspberry Pi. GitHub, but I can prepare prebuilt versions of.I am not able to publish the binarys due to the aseprite license, unless wants to publish them ARM / Raspberry Pi support I had to build my own version of gn, ninja and the skia “aseprite-m96” branch for “armhf” (32bit armv7) to make the build compatible with a standard Raspberry Pi Raspbian install. The application uses about 50mb of memory thus with a minimal X11 system aseprite + windomanager uses only 10% of the available memory of a Raspberry Pi 2 with 1gb of RAM. (it took 7gb of c++ sourcecode and generated objectfiles to link the final 19Mb aseprite binary )ĭuring the compilation i had to pass -latomic to the linker in order for the final aseprite executable to link. It is probably working on Pi Zero 2 as well (untested).Īseprite is not working on Raspberry Pi 1 and Raspberry Pi Zero 1 due to missing support for armv7 instructions on the CPU.Ĭompilation took about 2h on a Raspberry Pi 400 board. I have successfully compiled and tested aseprite 1.2.X and 1.3.X-beta with skia laf backend that is working well on the Raspberry Pi 2, 3, 4 and 400 model with OpenGL enabled. I have built the latest main on my Pinebook Pro (aarch64 CPU) using this method. If it’s not available, I’d try getting the binary from the Arch Linux ARM repos here. I’m not sure if the raspberry pi os repo’s have gn, but they should have ninja. Then I needed to symlink python2 to python, as that’s what the skia build script seems to expect.Īseprite built without that much extra stuff, but I had to set one library (I think libcurl) to shared, as it was throwing a few errors. To compile skia I needed to download ninja and gn for aarch64 instead of using the binaries provided in depot_tools. It’s true that more and more notebooks are in aarch64, I’ll check if we can publish a skia version for arch64 soon (at the moment linux is the lowest priority)Īnd while building Aseprite on aarch64 is not possible by following the official instructions, it’s not impossible. It would be nice to have a arch64 version, but it’s not high priority, it would be good for users with rasp pis, maybe one version each X years (not the latest one)
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |