I have to note next think:
Lets to see gxneur-devel-git - package has version 0.20.0-17, in the PKGBUILD file we can see
pkgname=gxneur-devel-git
pkgver=0.20.0
pkgrel=17
so installed application has the same version - 0.20.0-17 and package manager can see when source is updated. With PhF we have another situation - packfge in repo and installed application ALWAYS have different versions.
Again:
:: 1 Packages to upgrade.
1 aur/photoflow continuous.r1286.g6180401a-1 ā 0.2.8.r2.gdf03f253-1
As I think there isnāt any updates from yesterday but we have continuous.r1286.g6180401a-1installed and 0.2.8.r2.gdf03f253-1 in AUR.
Name change is somewhere in source files from git, PKGBUILD is correct, I guess.
Have you contacted the AUR packaging person to have them help see what the problem might be?
AUR packager creates valid PKGBUILD, but all files including makefiles are downloaded from git.
Yes I understand how the AUR works, you could still ask them, or at least confirm that the problem isnāt isolated to just you.
I donāt think that this is very popular application and have many installations.
I signal about problem in AUR, and saw there notification about to use photoflow-git.
In AUR there are two packages, one is photoflow and corresponds to the latest official release (0.2.8), which is now quite outdated, and the second one is photoflow-git, which corresponds to the current git head of the stable branch. This is the package I recommend to use for the time being, at least until I manage to prepare a new release.
As for why the photoflow-git package is still stuck at the git version from the 26th of June I do not really know, as I have no control over the AUR package. I have just posted a request to update the package, letās see if we get some quick feedbackā¦
Hi!
PhF-git was successfully installed on my working com yesterday, but today it not compiled for installing to mt home comp/
cc1plus: error: unrecognized command line option ā-Wno-deprecated-registerā [-Werror]
cc1plus: all warnings being treated as errors
make[2]: *** [src/external/lensfun/libs/lensfun/CMakeFiles/lensfun.dir/build.make:193: src/external/lensfun/libs/lensfun/CMakeFiles/lensfun.dir/mod-pc.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:876: src/external/lensfun/libs/lensfun/CMakeFiles/lensfun.dir/all] Error 2
make: *** [Makefile:152: all] Error 2
Are you in this case compiling the sources manually or through AUR? If manually, could you send me the cmake command and also the full compilation log?
The unrecognized command line option should not trigger a fatal error, unless there is some other compilation issue associated with it, which I cannot see from what you posted aboveā¦
Thanks!
Program is compiled by your cmake from AUR with -Werror flag.
Thanx to all!!! PhF-git compiled and installed!
Iāve actually not done anything yet, at least not today, but still I am glad to hear that the compilation went fine this time! Do not hesitate to drop a message if problems appear againā¦
Could you provide me the input RAW file as well as the PFI file that generates the corrupted image? I can have a look at it later this eveningā¦
Do you get the same result with the other sharpening methods?
The Richardson-Lucy sharpening method doesnāt handle negative input values well. You have a black point adjustment in the Basic Adjustments that makes the dark parts of your image < 0. If you add a āClip valuesā before the sharpening layer, the problems go away.
There seems to be a major topic split here from layout to other questions. @Carmelo_DrRaw @paperdigits and friends.
RawTherapee works correctly with changed black point and any sharping methods.
I have been side-tracked by lots of things recently, which slowed down the PhF development a little bit. But the software is still there to serve you!
By the way, one of the things that diverted my attention is the implementation of local laplacian filters, something I wanted to explore since quite some time. Darktable has an implementation of them, and I actually took part of the code from them, but I first wanted to test the implementation recommended in the original papers.
I am planning to post some results tomorrow, although an implementation into PhF is still not around the cornerā¦
