![]() usr/local/include/lensfun/lensfun.h:2378:16: note: candidate constructor not viable: requires 4 arguments, but 5 were providedĭEPRECATED lfModifier (const lfLens *lens, float crop, int width, int height) Mod = new lfModifier(d->crop, w, h, LF_PF_F32, (force_inverse) ? !d->inverse : d->inverse) Users/chans/Local/dev/darktable/src/iop/:346:13: fatal error: no matching constructor for initialization of 'lfModifier' In file included from /Users/chans/Local/dev/darktable/build/lib/darktable/plugins/introspection_:85: Does anyone know what might be going on here? Clicking the settings icon results in a crash with error (darktable:25626): GLib-GIO-ERROR **: 09:09:51.573: No GSettings schemas are installed on the system.Seems to be mostly working, even with opencl support! Pixel pipeline processing benchmarks are about 25% faster with the M1 build as opposed to the Intel build from GitHub releases. Manually compile and install openmp-13.0.0rc1 from (version 12 results in OMP: Error #131: Thread identifier invalid crashes).cmd_config="$cmake -build \"$BUILD_DIR\" -target install -j$MAKE_TASKS" I know it is splitting the attention, and I wish I could help (musician/photographer with some python experience), but it would be great to have darktable running on ARM mobile, and not have to just shop for small devices in the x86 eccosystem (thinking of getting a windows go tablet and putting linux on it for travel, but the idea kinda turns of stomach :)Īgain, sorry if the question is off subject, but, if the M1 build works, perhaps it can be done for other ARM devices? I'd volunteer as a test compile/install mule, but I'm afraid the programming side isn't my forte. (I know, it is a pipe dream to have darktable available on mobile through F-droid.) We now have 64 bit rasberry Pi's, and a lot of android is on 64 bit ARMs. Sorry if I am asking a question WAY out of my league, but what will this mean for building darktable for other 64 Bit ARM architectures? I know there was some interest in the past for running these on other ARM, but, even if it could be complied, it was all on 32 bit, and the claim was it would be crazy slow. So someone needs to backport the ARM64 fix to 0.3.2 release, or figure out the whole mess with lensfun missing a release and being incompatible with parameters from older version. I think homebrew has lensfun package working on M1, but it's version 0.3.95, which is buggy, so it's a no go. In any case I don't think it's a good idea to release a package without lensfun support. I haven't tested performance yet, I wouldn't be surprised if Rosetta-translated version of darktable actually runs faster than native build.Īnd an open question related to #8589 - do we want to have one universal DMG package for both Intel and ARM64 Mac platforms or make 2 separate DMGs (easier)? Merging 2 builds with some script calling lipo probably isn't hard, but from a practical point of view having an universal build robs a user from an option to use Rosetta-translated darktable, in case native build crashes or is just slower. On darktable side only -DBUILD_SSE2_CODEPATHS=OFF cmake option was needed to make build succeed. libavif - no such port in macports, so missing in my Intel build also, just an additional feature, probably not used/wanted by many.gtk-osx-application-gtk3 - just some GUI integration sugar, not critical, but nice to have.lensfun - even being optional, it's quite essential to have, since it's needed for an iop. ![]() ![]() Missing are the following optional dependencies, in order of importance: Also, new hardware support can be delayed depending on who is available to code and review it, but when that support does get committed, it is typically better than the manufacturers' own support.I've managed to compile and start darktable on M1. Yet the precision and versatility of this collection of software is well worth the effort, and helps to encourage better exposures to start with.īe aware that the core development occurs under Linux, the OS X version is a port maintained by a separate group of volunteers, and there often is a lag between versions. This requires a very clear concept of what one wants out of an image - not some vague sense of "vibrant" or "sharper" - which can be challenging at first, especially for one accustomed to dumbed-down editors that offer automated quick fixes. It is not intended to be an "editor" but rather a "finisher" of images. One of the best suites of tools for working with RAW image files regardless of price, and with its very active and cooperative developers community it just gets better and better. ![]()
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |