control: tags -1 -moreinfo Dear Paul,
Thanks for your checking! On Thu, May 27, 2021 at 5:50 PM Paul Gevers <elb...@debian.org> wrote: > > Control: tags -1 moreinfo > > Hi, > > On 17-05-2021 02:12, Ben Hutchings wrote: > > Please unblock package broadcom-sta > > > > [ Reason ] > > Fix the unusable broadcom-sta-source package. > > > > [ Impact ] > > It is not possible to build a package using module-assistant and the > > version of broadcom-sta-source in bullseye. However, dkms and > > broadcom-sta-dkms can be used as an alternative. > > > > [ Tests ] > > Only the build processes are being changed. I tested that: > > - broadcom-sta can be built from source > > - module-assistant can build a module package from broadcom-sta-source > > for the current kernel version in bullseye (5.10.0-6-amd64) > > - the resulting binary module package looks like a module package > > built from broadcom-sta-source in buster, modulo version changes > > * I wonder, broadcom-sta has seen quite some uploads the last couple of > years and debhelper is even in oldstable newer than the version > mentioned. How were all these uploads possible? I think "some uploads" means uploading "src:broadcom-sta", which states debhelper version in debian/control. And debian/control is not updated in this unblock request. The source updated in this upload is: debian/control.modules.in which is not used for upload, and will be explained below. > * What is/was the behavior of debhelper if the compat level was not > specified? In the freeze we don't want debhelper compat bumps unless the > package is proven to have no delta regardless of the old-vs-new level. The issue resolved in this upload is: after installing broadcom-sta-source, when user try to build their own deb files by using tool module-assistant, the deb build would fail. The user built deb is not for upload to debian archive, but for private use only. Personally I don't install and use broadcom-sta-source, so I didn't notice this issue for years. I hope things get clear now. Thank you! Cheers, -- Roger Shimizu, GMT +9 Tokyo PGP/GPG: 4096R/6C6ACD6417B3ACB1
