On Fri, Dec 01, 2023 at 12:24:12PM -0500, Patrick Robb wrote: > On Fri, Dec 1, 2023 at 11:36 AM Ferruh Yigit <[1]ferruh.yi...@amd.com> > wrote: > > On 12/1/2023 2:30 PM, Bruce Richardson wrote: > > On Fri, Dec 01, 2023 at 11:33:25AM +0000, Ferruh Yigit wrote: > >> On 12/1/2023 11:13 AM, Bruce Richardson wrote: > >>> On Fri, Dec 01, 2023 at 09:04:10AM +0100, Thomas Monjalon wrote: > >>>> 30/11/2023 19:33, Patrick Robb: > >>>>> On Thu, Nov 30, 2023 at 4:24 AM David Marchand > <[2]david.march...@redhat.com> > >>>>> wrote: > >>>>> > >>>>>> What it means: > >>>>>> - for the [3]https://dpdk.org/git/dpdk repository, all the > branches and > >>>>>> tags are mirrored to [4]https://github.com/DPDK/dpdk as it > was done so > >>>>>> far, > >>>>>> - for the [5]https://git.dpdk.org/next/dpdk-next-* > repositories, only > >>>>>> branches named "main", "staging" or "for-*" are mirrored to > >>>>>> [6]https://github.com/DPDK/dpdk with a prefix. > >>>>> > >>>>> Thank you David for clearing some of this up on the CI testing > meeting. I > >>>>> think the final loose end was you were wondering which > branches within the > >>>>> next-* repos we were running from. I'll paste that below: > >>>>> > >>>>> dpdk-next-crypto: for-main > >>>>> dpdk-next-eventdev: for-main > >>>>> dpdk-next-net: main > >>>>> dpdk-next-net-brcm: main > >>>>> dpdk-next-net-intel: main > >>>>> dpdk-next-net-mlx: main > >>>>> dpdk-next-net-mrvl: for-next-net > >>>>> dpdk-next-virtio: main > >>>>> dpdk-next-baseband: for-main > >>>> > >>>> We should test patches on top of the branch which is validated > >>>> by the tree maintainer and ready to pull. > >>>> This is the default branch (HEAD) of its repository on > [7]dpdk.org. > >>>> This is the list of equivalent GitHub branches to use for > testing: > >>>> > >>>> main > >>>> next-baseband-for-main > >>>> next-crypto-for-main > >>>> next-eventdev-for-main > >>>> next-net-for-main-repo > >>> > >>> The (slight) inconsistency here is curious. Is there a reason > why this > >>> branch has "repo" on the end and none of the others don't? > >>> > >> > >> No specific reason, it started like that in the past. If the > consensus > >> is to go with 'for-main', I can update it. > >> > > In all likelihood it's not a big deal. > > > > I'm just thinking that for any future automation or tracking, > having a very > > consistent naming would be best. For example, in the one-liner I > posted for > > generating the graph showing the inter-tree flow, I needed a > special regex > > to strip off the "-repo", so that we didn't have a separate target > called > > "main-repo" and another called "main". > > > Agree on above, created 'for-main' branch and will use it now. > But I need help from Thomas/David to delete remote 'for-main-repo' > branch. > > Great. So for CI testing, we will use dpdk-*-for-main when possible. > next-virtio only has for-next-net and staging, so we will use > for-next-net, unless we want to rename that one too to have a true > standard.
That one is correct, because next-virtio doesn't go straight to main. There will soon similarly be next-net-intel-for-next-net, for example, for the other sub-trees that get pulled into next-net. /Bruce
