<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Jan 16, 2020 at 11:00 AM Renato Golin <<a href="mailto:rengolin@gmail.com">rengolin@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Thu, 16 Jan 2020 at 18:45, David Blaikie <<a href="mailto:dblaikie@gmail.com" target="_blank">dblaikie@gmail.com</a>> wrote:<br>
> I'm not sure where the idea that a patch series is anything other than that ^ came from. When I was talking about a patch series, it was/is with that definition in mind - ordered/dependent commits. I said "dependent series" to reinforce this idea that the kind of situation I was describing was one in which later patches in the series depend on the changes in earlier patches.<br>
<br>
Perhaps it's my confusion in interpreting the answers.<br>
<br>
Some comments mentioned "if a review spawns a related change", which<br>
to me is a different review, and I've reviewed many of those cases.<br>
Most people use the Phab link to express relationship, but that's not<br>
a series.<br></blockquote><div><br>Can you point to examples of that - where Phab links have been used to express non-mechanically-dependent patches?<br> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Others said "some patches may be approved before others" </blockquote><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">which only<br>
works in a series if they're from start to N, not N to M, nor M to the<br>
end, nor randomly approved. In this case, the series is split in two,<br>
with the latter having to be rebased on patches committed after the<br>
first part is, so essentially, creating a new series.<br></blockquote><div><br>Approval order isn't commit order - I'm more than happy to approve a later patch in a series, even if the review of an earlier patch necessitates some rework in a later patch - such as renaming a function. The later patch I already approved must be updated to use the new name of the function, but I'm fine with that, same as I would be if it'd been an independent change that did the rename in the time it took us to review that patch.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Doing both these cases as a pull request is trivial.<br>
<br>
Related changes become separate PRs with mention. </blockquote><div><br>What do you mean by "with mention" and what do you mean by "related" (I guess you man "not dependent, but interesting to consider together")<br> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">GitHub creates the<br>
links, like Phab if you tag on the commit message or comments.<br>
<br>
Series split is harder, but still trivial. You create a new branch,<br>
move up to the approved patch, push. Rebase your old branch and you<br>
have a new series.<br></blockquote><div><br>My understanding is that this ^ is the case we're talking about with Phab patch series. Related but independent patches aren't "interesting", as you say - just mention them in passing in the review.<br><br>My understanding from other people's comments (I've not tried it myself) is that updating a patch series in github is problematic - that it sends new/separate review email rather than being able to associate each patch in the series with ongoing review feedback and updates?<br> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">In Github you can choose to close the PR and open a new one, or just<br>
push again and the UI should update the range. I prefer the former,<br>
because you keep the comments history.<br></blockquote><div><br>Perhaps there's some misunderstanding/different experiences there (as there is with Phab). Maybe you could make a little example? Showing how a dependent patch series with ongoing review feedback works with github PRs?<br> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<br>
I would have done the same process in Phab, but with more clicks,<br>
uploads, links, etc.<br>
<br>
--renato<br>
</blockquote></div></div>