Arch Linux User Repository

Turns out that the error is caused by uninstantiated template function read_volumeROI (see line 4841 of the log). This can be fixed by the following patch: https://gist.github.com/wangqr/ab9d47ec1d955310794256905d60d583

After fixing this issue, there are still 4 components that fails to build, namely fslsurface, film, ptx2, and MSM. The latter 2 depends on fslsurface. fslsurface fails due to corrupted .o file (not sure why this happens). film fails due to missing header QPBO/QPBO.hpp

The build fails for me. Full build.log at https://gist.github.com/wangqr/ab96316acbc92e640b6d044688b77648

It seems that the errors are caused by missing headers. According to the command line options (See line 4868 of the log), the compiler is looking for headers in $HOME/fsldev, which does not exist.

Thanks, I tried reinstalling fslpy and to run the test suite manually from the executable. Nonetheless the same output came out without any new information about the problem.

I think the issue might be a problem in the building itself. I will try to read through the build.log.

I ran the test suite manually and a more useful Python error came up. There should be an executable called RUN in the src/feeds directory. A couple of environment variables need setting up (which is fiddly), but you can find the required steps in the check() section of the PKGBUILD file.

I'd also try updating and reinstalling the fslpy package if you haven't done that already.

@martin3141 I'm having the same error with check() that you reported. It is the exact same output. I tried your solution of reinstalling the python-deprecation package, but the failure keeps occurring. How did you realize which package was missing from your system?

I'm glad it's working for you now. Thanks for mentioning the fslpy dependencies issue, I sadly failed to notice AKSoo's comment.

v6.0.1 is working for me now - thanks :)

My original problem with the check failure was also an issue with v6.0.1. Turned out the cause was that the python-deprecation package was missing from my system, maybe this should be added to the makedepends for fsl, or perhaps it should be a dependency (rather than a makedepends) for fslpy (there is a comment suggesting this from AKSoo).

Well, it's even simpler and somewhat embarrassing. I forgot to copy the adapted 001-use_distribution_environment.patch from my test environment to my "release environment". That's why it worked for me and didn't for you. You shouldn't do things in a hurry. Fixed.

Thanks for updating! I think you forgot to update a checksum : "001-use_distribution_environment.patch ... FAILED"

Skipping the check with --skipinteg reveals a further issue: " patching file fsl/etc/fslconf/fsl.sh patching file fsl/etc/fslconf/fsl.csh patching file fsl/etc/fslconf/fsl-devel.sh patching file fsl/extras/build patching file fsl/src/mist-clean/Makefile can't find file to patch at input line 110 Perhaps you used the wrong -p or --strip option? "

Updated to FSL version 6.0.1. I also decided to reduce the clunkiness of 001-use_distribution_environment.patch and ship {externallibs,systemvars}.mk as separate files. Seems less ugly this way.