-
Notifications
You must be signed in to change notification settings - Fork 18
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SIRF-SuperBuild build fails on Windows #197
Comments
It looks like the error is coming from SIRF (thus If so, it sounds like |
Here are the arguments in
It doesn't look like |
Thanks, it worked!
where
I hope this will not break anything - any comments anyone? |
I would have thought that you don't need to set any variables, since these would all be set in When I'm debugging, I print all available CMake variables like this:
If I were you, I would put this somewhere in your |
will try later now i have run into another problem when trying to SuperBuild with DEVEL_BUILD on:
BOOST_LIBRARYDIR was set as suggested (to Boost/bin.v2/libs) |
Not sure for Windows, but try setting
|
no, it does not help, same error message |
Well again, this looks like a SIRF error (not superbuild), so check the value of |
I never succeeded to get CMake to correctly find a Boost version that we compile. It's a very complicated moving target with boost renaming libraries in the build schemes, and CMake trying to catch-up. I very much recommend to install the Boost binaries, and set |
Regarding the missing Otherwise, I'm surprised as it searches for the library with find_library , see for example here. I thought this would go and search in our install directory., but I admit that I cannot see anything related to that in the |
See #7 for other remaining issues with the Windows build. Note however that I succesfully build on Windows 10 with CMake 3.13.2 with |
same problem with boost on windows 7 all boost libraries are actually present in moreover, there are no boost dlls in build/SIRF-SuperBuild at all |
Surely all the dlls belong in |
Please do not waste time trying to set |
no, it does not help, same error message |
installed boost_1_65_1, boost is ok, back to fftw3 problem:
cmake shows correct values for |
This commit makes little sense to me, given the discussion in #147. Maybe I was incorrect there. In any case,
Can we do this via a PR please? |
This re-closed itself automatically, not sure why. |
ah... Thanks! |
my standalone SIRF has CMake option |
Matlab fails with the error message
apparently there is a conflict between HDF5 libraries used by SuperBuild and matlab's ones |
overwriting matlab hdf5 dlls (luckily, there are only two of them) does fix the problem, but it is hardly a solution many users would like :( |
could prefix to matlab's path. Can also contact matlab to get them to update the dlls (I've done this before) |
This is the same problem as what we're having with Boost https://github.com/CCPPETMR/SIRF/issues/240. Like there, I think the best thing to do is to build SIRF with the same version as what is used by Matlab. That implies that we need to check what the HDF5 version is for different Matlab versions. Fun. @evgueni-ovtchinnikov would you be able to send an email to Mathworks support requesting the relevant information (for both HDF5 and Boost). Best to explain them why. feel free to point them to this issue. |
Proposed FFTW fix is in #204 |
I've tried this again with latest versions.
I'll update this when I find more |
I get warnings when building STIR via the SIRF-SuperBuild:
Clearly, |
resolved this with 75281a5 |
I've resolved all the above problems now. SIRF builds without any problems with the SIRF-SuperBuild! Sadly, all tests fail... |
Tests failed due to path issues. I also needed SyneRBI/SIRF#840. Setting the paths as in #460, all tests work!
You'll notice though that these were with MATLAB only as I currently don't have Python on my Windows laptop. I'm 99% sure that I will have to disable MATLAB when building with Python and vice versa due to HDF5 version conflicts etc, but at least things are looking up! |
I have now tested this with miniconda's Python 3.8. I've put in a small commits but it all works for me, although I do need .SyneRBI/SIRF#844 as well. Note that I did make a separate build for Python. |
So, I believe i can close this now. There's 2 SIRF PRs pending, but as far as the SuperBuild goes, all is well. Some info on my test environment (all very new):
CMake Configuration:
run-time environment was as in #460 I built the |
SIRF-SuperBuild build fails on Windows with the error message from CMake:
CMake GUI however displays both 'missing' items with correct values!
The text was updated successfully, but these errors were encountered: