On 28. 01. 22 3:37, Troy Dawson wrote:
On Wed, Jan 26, 2022 at 7:54 AM Troy Dawson <tdawson@redhat.com mailto:tdawson@redhat.com> wrote:
On Wed, Jan 26, 2022 at 7:37 AM Miro Hrončok <mhroncok@redhat.com <mailto:mhroncok@redhat.com>> wrote: On 26. 01. 22 16:30, Troy Dawson wrote: > EPEL 8 Playground is going away. > One of the steps in that process [1] is to clean out playground from all the > various package.cfg files. > I will not be removing the package.cfg files. I will only remove > epel8-playground entry if it is there. If you, as a package maintainer, want > to remove the package.cfg file, you are free to do so. > I have seen too many package.cfg files that have been modified, and I do not > feel safe globally removing them. > > Note: I will be checking the epel8, epel9, rawhide and f35 branches for > package.cfg files. > > This will be happening later today. Let me know if you have any concerns > and/or comments. Hey Troy. Could you please share the list for inspection before actually changing anything? I can, and will. Good idea. It will be a couple hours before I have that list. Troy
That took longer than expected. Sorry about that. I know I said that I was only going to take the epel8-playground out of the files, but it turned out that there were so many that only have the default package.cfg that we put in, that I feel we should take all those default files out. There was three groups.
** A - Custom package.cfg
- I will only remove epel8-playground
argbash (custom) rawhide f35 nss-mdns (custom) f35 RBTools (custom) rawhide f35
** B - Default epel8 package.cfg - in Rawhide and F35
- I am going to remove the package.cfg from rawhide and f35
beanstalk-client (default) rawhide f35 copr-selinux (default) rawhide f35 czmq (default) rawhide f35 fctxpd (default) rawhide f35 gedit-plugin-editorconfig (default) f35 glances (default) rawhide f35 gnome-doc-utils (default) rawhide f35 libwebsockets (default) rawhide f35 MUMPS (default) f35 netcdf4-python (default) rawhide f35 opentrep (default) rawhide f35 python-astroid (default) rawhide f35 python-cftime (default) rawhide f35 python-kubernetes (default) rawhide f35 python-lazy-object-proxy (default) rawhide f35 python-multidict (default) rawhide f35 python-repoze-tm2 (default) rawhide f35 python-repoze-who (default) rawhide f35 python-transaction (default) rawhide f35 R (default) f35 sagator (default) f35 TurboGears2 (default) rawhide f35
...
Let me know if anyone disagrees with my plan.
Thank You!
Looking for example at python-astroid where your plan is to remove it from f35 and rawhide but not from f34.
The f35 and rawhide branches are not in sync but f35 is "reachable" from rawhide history. Do we really need to diverge f35 just to remove a file that we are OK keeping on f34?
Looking at python-astroid in Koji: https://koji.fedoraproject.org/koji/packageinfo?packageID=16809
It doesn't seem this was submitted regularly for many targets. The file has:
[koji] targets = epel8 epel8-playground
Yet when I run `fedpkg build` on rawhide, it only submits a build for rawhide. Similarly on f35, it only submits a build for f35.
When I run `$ fedpkg --release=epel8 build` it submits 2 builds, so it indeed does at least something. The dangerousnes of this is... minimal? Considering the Koji target will be blocked.
Hence I propose to only remove the file from f35 if the branch has the same HEAD as rawhide, but not to remove it otherwise to avoid git mess.
Similarly, I would also remove it from f34 in such case.