Hello golang-codereviews@googlegroups.com (cc: adg@golang.org, dvyukov@google.com, rsc@golang.org), I'd like you to review this change to https://code.google.com/p/go.tools
9 years, 7 months ago
(2014-09-14 18:32:45 UTC)
#1
LGTM I guess https://codereview.appspot.com/136540044/diff/40001/dashboard/builder/vcs.go File dashboard/builder/vcs.go (right): https://codereview.appspot.com/136540044/diff/40001/dashboard/builder/vcs.go#newcode213 dashboard/builder/vcs.go:213: <ChangeNum>{rev}</ChangeNum> I thought this value was ...
9 years, 7 months ago
(2014-09-15 00:29:56 UTC)
#3
In theory, but for a builder with a fresh hg clone checkout and no local ...
9 years, 7 months ago
(2014-09-15 00:32:02 UTC)
#4
In theory, but for a builder with a fresh hg clone checkout and no local
commits (not that we've ever used hg that way anyway), I think ChangeNum is
predictable.
On Sun, Sep 14, 2014 at 8:29 PM, <adg@golang.org> wrote:
> LGTM I guess
>
>
> https://codereview.appspot.com/136540044/diff/40001/
> dashboard/builder/vcs.go
> File dashboard/builder/vcs.go (right):
>
> https://codereview.appspot.com/136540044/diff/40001/
> dashboard/builder/vcs.go#newcode213
> dashboard/builder/vcs.go:213: <ChangeNum>{rev}</ChangeNum>
> I thought this value was essentially useless, as different checkouts can
> provide different numbers.
>
> https://codereview.appspot.com/136540044/
>
If you do a fresh checkout every time, maybe. But I think it also depends ...
9 years, 7 months ago
(2014-09-15 00:41:33 UTC)
#5
If you do a fresh checkout every time, maybe. But I think it also depends
on the order in which the hg server sends you the changes.
As soon as you do "hg pull" all bets are off, though.
https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handler.go File dashboard/app/build/handler.go (right): https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handler.go#newcode150 dashboard/app/build/handler.go:150: if com.ParentHash != "" { I think the original ...
9 years, 7 months ago
(2014-09-18 00:00:50 UTC)
#6
LGTM https://codereview.appspot.com/136540044/diff/60001/dashboard/builder/vcs.go File dashboard/builder/vcs.go (right): https://codereview.appspot.com/136540044/diff/60001/dashboard/builder/vcs.go#newcode205 dashboard/builder/vcs.go:205: <Parent>{p1node}</Parent> split this change into a separate cl?
9 years, 7 months ago
(2014-09-23 00:30:10 UTC)
#9
https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handler.go File dashboard/app/build/handler.go (right): https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handler.go#newcode150 dashboard/app/build/handler.go:150: if com.ParentHash != "" { On 2014/09/18 00:00:49, adg ...
9 years, 7 months ago
(2014-09-23 18:30:52 UTC)
#10
https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handl...
File dashboard/app/build/handler.go (right):
https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handl...
dashboard/app/build/handler.go:150: if com.ParentHash != "" {
On 2014/09/18 00:00:49, adg wrote:
> I think the original test is fine. This test is so that we don't look for a
> previous commit if this is the first commit we have seen. We don't actually
want
> the dashboard to build from the very first commit.
I disagree. This line was giving too much meaning to the "Num" field. The
comment says it's about testing the parent, but then it goes to use the sort
order.
https://codereview.appspot.com/136540044/diff/40001/dashboard/app/build/handl...
dashboard/app/build/handler.go:163: if p.Path == "" &&
!strings.HasPrefix(com.Desc, "[release-branch") && !isUpdate {
On 2014/09/18 00:00:49, adg wrote:
> we should probably change the prefix test to just "[" so that it catches the
new
> branches too
Done.
Issue 136540044: code review 136540044: dashboard: more robust commit handler, and allow modifi...
(Closed)
Created 9 years, 7 months ago by bradfitz
Modified 9 years, 7 months ago
Reviewers:
Base URL:
Comments: 10