Hi Andreas > > I have already updated the GitHub repository. I updated the xxhash and > > robin-map libraries to their latest versions and adjusted the source code > > to avoid issues. > > Very good. This kind of cooperation with upstream is IMHO a very > valuable (and perfectly intended) effect of the Debian Med project. > I've adapted the packaging to use the Debian packaged libs.
Perfect. > > Now we can remove them in Debian and use the packages. I > > also fixed the linking error in the CMakeLists file and, by the way, > > allowed default values in CMAKE_CXX_FLAGS, which I suppose was causing the > > blhc failure. If you believe that there is nothing else to add, tell me, > > and I will proceed with creating the 4.0.1 release. > > Sounds pretty much a sensible point to add a new release tag. However, > there is some issue with the CI test you suggested. I followed your > suggestion to do the very same test as for fasttree[1]. Unfortunately > this results in > > TopHits: 1.00*sqrtN close=default refresh=0.80 > ML Model: Jones-Taylor-Thorton, CAT approximation with 20 rate categories > Error parsing header line: > > as you can see in the log on salsa[2]. Ohh, well, the test has done its job, it found a bug. I have already fixed it, so is everything ready for the release now? > > In the Salsa repository, I tried to modify the rules to ignore AVX2 with > > ARM, but for some reason, it doesn't seem to recognize the DEB_BUILD_ARCH > > variable and always uses it. Did I forget something?. I haven't tested it, > > but otherwise, the source code should be compatible with ARM. > > I did not checked (yet) but you might like to have a look into > SIMDEverywhere[3] Thank you, but I meant that I was enabling AMD64 features on ARM. I've taken a look at the log and I see what the problem might be. Considering that aarch64-linux-gnu-g++ is being used, it appears to be a compilation for ARM, but with BUILD_ARCH=amd64 and HOST_ARCH=arm64, shouldn't it be the other way around? In any case, if this isn't an error, I would look for when HOST_ARCH=arm64 and wouldn't use USE_AVX2=ON, which is the parameter that sends -mavx2. Because right now, I'm using BUILD_ARCH, and since it's amd64, it assumes it should use USE_AVX2=ON. César
