On 25.01.2019 21:56, Ben Pfaff wrote: > On Fri, Jan 25, 2019 at 04:08:10PM +0300, Ilya Maximets wrote: >> On 23.01.2019 15:40, Aaron Conole wrote: >>> Ilya Maximets <[email protected]> writes: >>> >>>> Aaron, what do you think about supporting tags like 'Based-on' in ovsrobot >>>> ? >>>> For example, patchew in QEMU parses 'Based-on: <Message-Id>' tags. >>>> Maybe we can use patchwork ids or something similar. >>> >>> I think there's a way to do it without a tag in the commit message. I >>> don't like having a tag for this sort of information in the commit >>> message itself since it isn't useful to preserve in the commit history >>> (and can serve to clutter things up). >> >> I don't like the idea having such tags in git log too. But we could place >> them under the cut line like in this patch. But, yes, developers will >> frequently misplace them in the main part of commit-message. OTOH, we could >> add a quick checkpatch rule for that. >> >>> >>> Currently, we use [PATCH branch-XXX] to imply that the series needs to >>> be based on a specific branch. The ovsrobot automatically creates >>> branches based on patch series. Maybe it makes sense to enhance that so >>> that some subject metadata marker (series-[0-9]+) would automatically >>> check out the series and then apply the patches. This also lets a >>> reviewer know at a glance that the patches are destined to be applied >>> after a specific series which is already posted. One downside is that >>> it's more subject line data. I can't think of others at the moment, but >>> I'm sure there are some. Since the robot doesn't currently clean up >>> branches, it's entirely possible to even base a series on a severely old >>> branch (so that should get fixed at some point). >>> >>> WDYT? >> >> Long subject lines is not a very good thing too. They hiding the patch >> subject in the mess of tags. >> Not sure what is better. >> >> Ben, do you have any thoughts about this? > > I'm not a big fan of patch series that have prerequisite patches. Is > there a common, good reason for them? I usually encourage people to > post self-contained series.
I'm not a big fan too. But it happens from time to time if the development is faster than acceptance and the patches are logically not connected, but depends on each other by the code overlapping. This, probably, not the big issue for OVS mail-list as it's not so active in compare with qemu. So, I'm not sure if we need to support any type of 'based-on' tagging. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
