Am Thu, 30 Oct 2014 19:34:17 +0100 schrieb gregor herrmann <gre...@debian.org>: > Just out of curiosity: > > The test failure in #759933 was fixed by a patch that changed the > dependencies in 0.8-1.1 from the versioned|general vresions to the > versioned ones: > > - llvm-3.4-dev | llvm-dev (>= 1:3.4), > - clang-3.4 | clang (>= 1:3.4) | clang (>= 3.4), > - clang-3.4 | clang (>= 1:3.4) | clang (<< 1:0), > - libclang-3.4-dev | libclang-dev (>= 1:3.4) | libclang-dev (>= 3.4), > - libclang-3.4-dev | libclang-dev (>= 1:3.4) | libclang-dev (<< 1:0), > + llvm-dev (>= 1:3.4), > + clang (>= 1:3.4), > + libclang-dev (>= 1:3.4), > > And now with those dependencies we see #764930 and the proposed patch > is change the dependencies back to the versioned variants: > > - llvm-dev (>= 1:3.4), > - clang (>= 1:3.4), > - libclang-dev (>= 1:3.4), > + llvm-3.4-dev, > + clang-3.4, > + libclang-3.4-dev, > > > The latter makes sense from the explanation; I'm just wondering where > the build failure in late August with exactly those dependencies came > from and why the change fixed them?
As stated in the log, posted in the bugreport, during compilation clang was called but could not be found: > cd /«PKGBUILDDIR»/obj-x86_64-linux-gnu/backend/src && clang -cc1 -x > cl -triple spir -ffp-contract=off -fno-builtin > -DGEN7_SAMPLER_CLAMP_BORDER_WORKAROUND --relocatable-pch -emit-pch > -isysroot /«PKGBUILDDIR»/obj-x86_64-linux-gnu/backend/src > /«PKGBUILDDIR»/obj-x86_64-linux-gnu/backend/src/usr/lib/beignet/ocl_stdlib.h > -o > /«PKGBUILDDIR»/obj-x86_64-linux-gnu/backend/src/usr/lib/beignet/ocl_stdlib.h.pch > /bin/sh: > 1: clang: not found Only the clang metapackage but not the clang-some.version package provides /usr/bin/clang. Hence, changing the dependency to the clang metapackage fixed the issue at hand (but in hindsight it was not the right fix (tm)). Cheers, Julian
pgpVzHfZB9CVs.pgp
Description: Digitale Signatur von OpenPGP
