Re: Hamlib 4.0 in experimental
Initially I did try a source only upload that was rejected with the following advice
> qsstv_9.4.4-2.dsc: Refers to non-existing file 'qsstv_9.4.4.orig.tar.xz'
> Perhaps you need to include the file in your upload?
> If the orig tarball is missing, the -sa flag for dpkg-buildpackage will be your friend.
I think I see what’s happened - I’ve followed through with the dpkg-buildpackage -sa and dput the _amd64.changes file, not the .dsc or a _source.changes file.
I’ll fix it later on with a source-only upload.
DH
--
Hibby
MM0RFN
On Thu, 28 Jan 2021, at 9:43 PM, Christoph Berg wrote:
> Re: Dave Hibberd
>> I built and tested using my normal git-pbuilder method, which would have used Hamlib4, but I wasn’t able to upload the resultant .changes, I guess as it was binary only. Advice in the email was:
>>> If the orig tarball is missing, the -sa flag for dpkg-buildpackage will be your friend.
>
> There are several things that look similar but are actually
> orthogonal:
>
> #1 The upload can include a binary .deb package (required for NEW
> uploads; not for DMs)
> #2 The upload can include a source .dsc package (the normal case for
> maintainer uploads)
> #3 If the source includes a .dsc, it can include a .orig.tar.* file
> (the default is to include it with a -1 upload)
>
> The "normal" dpkg-buildpackage build will result in #1 + #2
>
> But these days you should really upload source-only (with or without #3)
> since otherwise the package won't migrate to testing:
> dpkg-buildpackage -S
>
> "source-only" doesn't imply that there is an orig tarball included in
> the upload, that is only required if the tarball is not yet in the
> archive. But even if you aren't including it, the .dsc file uploaded
> needs to have to correct checksums for the tarballs in the archive.
>
> Christoph
>
Reply to: