Re: rocm-device-libs_4.5.2-1~exp1_multi.changes is NEW
Hi Étienne,
Oops, I forgot to commit some of my local changes to the repo.
Actually that build failure is trivial since I've reverted
the binary package name libhsa-runtime64-dev back to
libhsa-runtime-dev in order to go through NEW again.
That reminded me ... that this package does not have to pass
the NEW queue. But anyway I think I'm uploading it to experimental
even if with unsatisfiable dependencies. In this way we can
strive towards its reverse dependencies.
Will check and upload sometime later this week.
On Mon, 2022-01-24 at 22:16 +0100, Étienne Mollier wrote:
> Hi Mo,
>
> M. Zhou, on 2022-01-24:
> > Excellent. Now that all the dependencies for rocr-runtime are
> > uploaded.
> > Shall I upload rocr-runtime to NEW as well? Or any of you are eager
> > to do that?
>
> Thanks for your time on the rocr-runtime package! I don't have
> the eagerness to upload myself, as I haven't spent much time on
> it, but would be okay to review and proceed if you're caught by
> time this week.
>
> > The current master branch for rocr-runtime can build cleanly with
> > sbuild (with some --extra-packages flags). I can upload upon
> > confirmation of its readiness.
>
> Hmn, build attempt on my end with reprepro'ed rocm-cmake and
> rocm-device-libs as uploaded to NEW fails with dh_missing
> reporting uninstalled files:
>
> -------------------------8<-------------------------------------
> dh_missing -O-Dsrc -O-Scmake
> dh_missing: warning: usr/include/hsa/Brig.h exists in debian/tmp but
> is not installed to anywhere
> dh_missing: warning: usr/include/hsa/amd_hsa_common.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/amd_hsa_elf.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/amd_hsa_kernel_code.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/amd_hsa_queue.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/amd_hsa_signal.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa.h exists in debian/tmp but
> is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_api_trace.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_ext_amd.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_ext_finalize.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_ext_image.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_ven_amd_aqlprofile.h exists
> in debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/include/hsa/hsa_ven_amd_loader.h exists in
> debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/lib/cmake/hsa-runtime64/hsa-runtime64-
> config-version.cmake exists in debian/tmp but is not installed to
> anywhere
> dh_missing: warning: usr/lib/cmake/hsa-runtime64/hsa-runtime64-
> config.cmake exists in debian/tmp but is not installed to anywhere
> dh_missing: warning: usr/lib/cmake/hsa-runtime64/hsa-
> runtime64Targets-none.cmake exists in debian/tmp but is not installed
> to anywhere
> dh_missing: warning: usr/lib/cmake/hsa-runtime64/hsa-
> runtime64Targets.cmake exists in debian/tmp but is not installed to
> anywhere
> dh_missing: error: missing files, aborting
> The following debhelper tools have reported what they
> installed (with files per package)
> * dh_install: libhsa-runtime-dev (0), libhsa-runtime64-1 (2)
> * dh_installdocs: libhsa-runtime-dev (0), libhsa-runtime64-1
> (0)
> If the missing files are installed by another tool, please
> file a bug against it.
> When filing the report, if the tool is not part of debhelper
> itself, please reference the
> "Logging helpers and dh_missing" section from the
> "PROGRAMMING" guide for debhelper (10.6.3+).
> (in the debhelper package:
> /usr/share/doc/debhelper/PROGRAMMING.gz)
> Be sure to test with dpkg-buildpackage -A/-B as the results
> may vary when only a subset is built
> If the omission is intentional or no other helper can take
> care of this consider adding the
> paths to debian/not-installed.
> -------------------------8<-------------------------------------
>
> On first sight it looks like side effect of debhelper compat
> version bump to version 13, but it could as well be a side
> effect of my lastest changes to rocm-device-libs. The most
> relevant recent change I can think of in rocm-device-libs is the
> move of the directory /usr/amdgcn to /usr/share/amdgcn, for
> compliance with the file system hierachy. (Or I'm not looking
> at the right source.)
>
> Have a nice day, :)
Reply to: