Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(75)

Issue 22852017: git: make fetch chromium use unmanaged mode (Closed)

Created:
7 years, 4 months ago by Paweł Hajdan Jr.
Modified:
7 years, 4 months ago
Reviewers:
Dirk Pranke, szager, szager1
CC:
chromium-reviews, cmp-cc_chromium.org, M-A Ruel, iannucci+depot_tools_chromium.org, ilevy-cc_chromium.org
Base URL:
https://chromium.googlesource.com/chromium/tools/depot_tools.git@master
Visibility:
Public.

Description

git: make fetch chromium use unmanaged mode This has been determined to work in a better, more predictable and robust way. BUG=none R=szager@google.com Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=219121

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+1 line, -1 line) Patch
M recipes/chromium.py View 1 chunk +1 line, -1 line 0 comments Download

Messages

Total messages: 7 (0 generated)
Paweł Hajdan Jr.
7 years, 4 months ago (2013-08-22 21:32:55 UTC) #1
szager
lgtm * 1000
7 years, 4 months ago (2013-08-22 21:46:34 UTC) #2
Paweł Hajdan Jr.
Committed patchset #1 manually as r219121.
7 years, 4 months ago (2013-08-22 22:05:57 UTC) #3
Dirk Pranke
On 2013/08/22 22:05:57, Paweł Hajdan Jr. wrote: > Committed patchset #1 manually as r219121. Argh. ...
7 years, 4 months ago (2013-08-22 22:17:24 UTC) #4
Dirk Pranke
Also, I actually prefer managed to unmanaged and somewhat dispute the claim in the description. ...
7 years, 4 months ago (2013-08-22 22:19:47 UTC) #5
szager
On Thu, Aug 22, 2013 at 3:19 PM, <dpranke@chromium.org> wrote: > Also, I actually prefer ...
7 years, 4 months ago (2013-08-22 22:26:02 UTC) #6
Dirk Pranke
7 years, 4 months ago (2013-08-22 22:38:42 UTC) #7
On Thu, Aug 22, 2013 at 3:26 PM, Stefan Zager <szager@google.com> wrote:

> On Thu, Aug 22, 2013 at 3:19 PM, <dpranke@chromium.org> wrote:
>
>> Also, I actually prefer managed to unmanaged and somewhat dispute the
>> claim in
>> the description.
>>
>> I certainly don't think you should just land something like this without a
>> reference to a bug number or an actual description of why this is
>> changing;
>> otherwise this comment is indistinguishable from "I liked this default
>> better
>> than the other one".
>>
>> Note that I'm not opposed to this change if people do actually agree that
>> we
>> should flip this, but I want buyin from chromium-dev first, or at least
>> asking
>> of forgiveness. Otherwise, anyone who sets up a new checkout may be
>> confused
>> that src/ isn't actually being updated when they run 'gclient sync'.
>>
>>
https://codereview.chromium.**org/22852017/<https://codereview.chromium.org/2...
>>
>
> FTR, this has been discussed ad nauseum within the infra team, and I have
> had endless conversations will devs about it.  I can say with confidence
> that the tide of opinion is strongly against Managed mode.
>
> If you would like to throw it open to the teeming troll hordes on
> chromium-dev, I will head for the bomb shelter, and may god have mercy on
> your soul.
>

I'm fine w/ leaving the change in and seeing who gets upset by it, but I
strongly feel you need to at least tell chromium-dev@ about the change.

As I just said to Pawel in person, I recognize that it is very easy to
shoot yourself in the foot w/ managed mode. It is also easy to shoot
yourself in the foot w/ unmanaged mode.

I personally think the right thing to do is actually to have a third set of
behavior where gclient is smart enough to figure out what branch you're
tracking (if any) and either figure out what to do (merge/rebase/abort) or
prompt you. For the perhaps not-too-uncommon case of "I have no local
changes and am tracking origin/master", I think gclient sync should update
your checkout for you (just like it did in svn).

Does that make more sense?

In fact there's a bug on file for this:
https://code.google.com/p/chromium/issues/detail?id=228891 ... the last
time we had this discussion, szager noted a couple things, I commented on
them, and szager never replied. I kind of feel like we're doing and end-run
around that discussion now, rather than actually having it :).

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

Powered by Google App Engine
This is Rietveld 408576698