[aarch64-port-dev ] [RFR] 8u191 Upstream Sync

[Andrew Hughes]

On Wed, 2 Jan 2019 at 12:01, Andrew Haley <aph at redhat.com> wrote:
>
> On 12/20/18 4:31 AM, Andrew Hughes wrote:
>
> > This webrev brings jdk8u191-b12 (the upstream version) into the
> > aarch64/shenandoah-jdk8u repositories, tagging the result as
> > aarch64-shenandoah-jdk8u191-b13.
> >
> > https://cr.openjdk.java.net/~andrew/shenandoah-8/u191.upstream/
>
> This is OK. One thing, though: there are empty hunks like
>
>  Cdiffs Udiffs Sdiffs Frames Old New ----- Raw agent/src/os/linux/ps_core.c
>
>     0 lines changed: 0 ins; 0 del; 0 mod; 1092 unchg
>
> Cdiffs Udiffs Sdiffs Frames Old New ----- Raw agent/src/os/linux/ps_proc.c
>
>     0 lines changed: 0 ins; 0 del; 0 mod; 386 unchg
>
> Cdiffs Udiffs Sdiffs Frames Old New ----- Raw agent/src/os/linux/salibelf.c
>
>     0 lines changed: 0 ins; 0 del; 0 mod; 126 unchg
>
> Cdiffs Udiffs Sdiffs Frames Old New ----- Raw agent/src/os/linux/symtab.c
>
>     0 lines changed: 0 ins; 0 del; 0 mod; 545 unchg
>
> And also the entire AArch64 port seems to have been added. So what are these diffs?
> Is it that the AArch64-specific changes have been merged into the 8u191 upstream
> port, rather than upstream being merged into our repo?
>
> --
> Andrew Haley
> Java Platform Lead Engineer
> Red Hat UK Ltd. <https://www.redhat.com>
> EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671

I think the webrev tool gets very confused with the merge changeset for these,
which has two parents.
Upstream has been merged into our tree, but the merge changeset is effectively
joining together upstream and our local AArch64 changes, so it shows all the
AArch64 changes as new when compared with the upstream parent:

The merge for the HotSpot tree looks like:

changeset:   10721:bc4deb768b1d
tag:         aarch64-shenandoah-jdk8u191-b13
parent:      10652:e333239d5e9c
parent:      10720:4fc288749a23
user:        andrew
date:        Wed Dec 05 06:14:20 2018 +0000
summary:     Merge jdk8u191-b12

a change with two parents.

One parent is our last change:

changeset:   10652:e333239d5e9c
user:        aph
date:        Tue Nov 13 11:21:32 2018 -0500
summary:     8209415: Fix JVMTI test failure HS202

while the other is upstream:

changeset:   10720:4fc288749a23
user:        diazhou
date:        Wed Oct 03 21:50:07 2018 -0700
summary:     Added tag jdk8u191-b11 for changeset 12e4de4b2499

The difference between 4fc288749a23 and bc4deb768b1d is the whole of
AArch64 & Shenandoah i.e. anything not upstream.
This is why I include the actual merge changeset itself, which shows
the true difference:

https://cr.openjdk.java.net/~andrew/shenandoah-8/u191.upstream/hotspot/merge.changeset

It may make sense to just leave out the webrev in future because it
takes a long time to generate and isn't very useful.

Hope that helps,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Web Site: http://fuseyism.com
Twitter: https://twitter.com/gnu_andrew_java
PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222