When to do a review for a downported change?

Lindenmaier, Goetz goetz.lindenmaier at sap.com
Wed May 15 07:16:13 UTC 2019


This all sounds well reasonable and I agree to it.
To subsubme:
 - For pure hunk context changes and Copyright changes
   mention them in the "Fix Request" comment but don't
   send a downport RFR.
- If only a smallish change is needed to resolve
  supply a diff or other appropriate format in the mail.

Best regards,

> -----Original Message-----
> From: Aleksey Shipilev <shade at redhat.com>
> Sent: Tuesday, May 14, 2019 6:22 PM
> To: Andrew Haley <aph at redhat.com>; Lindenmaier, Goetz
> <goetz.lindenmaier at sap.com>; jdk-updates-dev at openjdk.java.net
> Subject: Re: When to do a review for a downported change?
> On 5/14/19 5:52 PM, Andrew Haley wrote:
> > On 5/14/19 4:02 PM, Aleksey Shipilev wrote:
> >> Aside, I think it is a good style (though optional) to post the diff
> >> between the upstream patch and the backport -- it seems low-overhead
> >> when there is the mq patch on top. This would also make reviews much
> >> easier, and probably fits the backporting workflow too. That is what
> >> I do anyway.
> >
> > Post in what format, though? A diff of a diff?
> Depends. Whatever fits the workflow, and maybe with workflow
> adjustments:
>  a) Sometimes just copy-paste *.rej and explain why those are not applicable.
>  b) Sometimes just inline the "addon" patches that fix the original change: in
> my workflow with mq
> there are two patches: the original "backport" change with rejects, and the
> follow up patch that
> resolves those rejects, hg qdiff the second one and paste it.
>  c) Sometimes just the copy-pasted affected block "before" / "after",
> showing the adjustment made;
>  d) Sometimes just the diff of a diff is enough to highlight the changes. If that
> is not available,
> I would do that during review for a non-trivial backport _anyway_.
>  e) Sometimes just pointing out which files and methods have differences vs
> upstream version, so we
> can eyeball them.
> I mean, anything that _you_ as reviewer, who sees the backport for the first
> time, would like to see
> in order to understand what was done, quickly and exactly. If backport
> differences are indeed
> trivial, then any of the ways above would amount to a copy-pasted
> paragraph in RFR, and it would
> breeze through the review, making the whole thing low-overhead.
> (This is not to say people here are not doing that. Just describing "the
> common sense" out loud.)
> -Aleksey

More information about the jdk-updates-dev mailing list