magma-rocm stuck in NEW queue?
Hello FTP Masters,
The magma-rocm package has been in the NEW queue for quite a
while [1]. Was there something about this upload that was of
particular concern?
The magma-rocm (2.9.0+ds-2) package is generated from the same
upstream tarball as magma (2.9.0+ds-2), so I'd expected it to
be a straightforward package from a NEW processing standpoint.
However, there is no formal mechanism that clearly expresses that
relationship, so perhaps it is not as useful to FTP Master review
as I had hoped? The Debian Science and Debian Deep Learning teams
have a number of source packages that share the same upstream
tarball for ROCm and CUDA variants, so perhaps we need to create
a mechanism to express that property in the packaging? Or maybe
we need to use a different pattern for packaging these variants?
I am really just speculating as to why magma-rocm seems to have
been a problem. Your guidance would be appreciated.
Sincerely,
Cory Bloor
[1]: https://ftp-master.debian.org/new/magma-rocm_2.9.0+ds-2.html
Reply to: