At the request of the RISC-V developers we are rebasing from libffi-3.4.2 to libffi-3.4.4. In order to test the rebase we used the Fedora Packaing tools mass-prebuild tooling to rebuild 144 packages (direct reverse dependencies of libffi) and evaluate the risk. There were 8 package build failures that needed analysis. Of the failures only 1 seemed related, Chromium, which failed to build with the new libffi. The others: bustle, llvm[9.0,10,11,12], p11-kit and ruby failed in the A-side and B-side tests. The Chromium build timed out when building in COPR, and then passed the build without a timeout on the A-side test (without libffi). This failure seems spurious and so I am moving forward with the rebase.
1 line
160 B
Text
1 line
160 B
Text
SHA512 (libffi-3.4.4.tar.gz) = 88680aeb0fa0dc0319e5cd2ba45b4b5a340bc9b4bcf20b1e0613b39cd898f177a3863aa94034d8e23a7f6f44d858a53dcd36d1bb8dee13b751ef814224061889
|