[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: rocsparse/5.3.0+dfsg-1 [ITP] -- ROCm library for sparse linear algebra



Thanks Étienne,

On 1/8/23 09:46, Étienne Mollier wrote:
Apparently, the last gbp import-orig didn't exclude files
specified in d/copyright, due to the statement not being in the
header paragraph.  Since the statement is now correct, you are
just one gbp import-orig away to fix this problem.  This is a
lintian error, so needs being addressed before upload.

I didn't know that import-orig used d/copyright, so my original import command was

gbp import-orig --filter=deps/convert \
    --filter=scripts/performance/matrices/convert \
    --filter=docs/source/_static/theme_overrides.css \
    --pristine-tar  --uscan --upstream-version=5.3.0+dfsg

It seems that command only removed the files from the master and upstream branches. While I was developing the package, I saw a number of lintian warnings and errors at the end of gbp buildpackage, but I addressed them all before uploading. If there were still lintian errors with the uploaded package, it sounds like there's something missing from my workflow. In any case, I've reimported and the files have now been removed from the (not so) pristine tar.

All the problem files have been removed upstream in the release/rocm-rel-5.5 branch, so we should be able to drop the +dfsg suffix in the next package update.

  * d/control:
	minor nitpick: the standards version has been bumped to
	4.6.2 recently, always good to keep the package up to
	date with the latest version, although changes since
	4.6.1 may not be too relevant for rocsparse.

Done. Although, this introduces the lintian warning

W: rocsparse source: newer-standards-version 4.6.2 (current is 4.6.1.0)
  * d/copyright:
	You might want to attribute yourself the copyright for
	2023 too.  :)
Done.
	I'm not entirely sure the mention of the Boost license
	is needed, since clients/benchmarks/program_options.hpp
	seems to be a form of reimplementation, but in doubt I
	suppose it can stay.  Besides, lintian informs us the
	paragraph is unused; this is a machine complaint from a
	machine readability expectation, but this is not a
	blocking problem.

Done. Though, how do you get these lintian warnings? I didn't see that warning even with lintian --pedantic.

I've reuploaded the package to https://mentors.debian.net/package/rocsparse/.

Sincerely,
Cory Bloor


Reply to: