Pin dEQP to ANGLE's DEPS file.

Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
Committed: https://crrev.com/60783171abc3fc122aeefd9a3705193e9191ccd9
Cr-Commit-Position: refs/heads/master@{#389262}

[Jamie Madill, 2016-04-22 19:48:13]

Description was changed from

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
==========

to

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
==========

[Jamie Madill, 2016-04-22 19:48:13]

jmadill@chromium.org changed reviewers:
+ kbr@chromium.org, maruel@chromium.org

[Jamie Madill, 2016-04-22 19:48:31]

Ken, MA, PTAL. Seems to be compiling and running the tests on the bots.

[Ken Russell (switch to Gerrit), 2016-04-22 20:14:46]

lgtm conceptually. I don't know this DEPS syntax.

[M-A Ruel, 2016-04-22 20:20:08]

yep, this sound work. lgtm

[Jamie Madill, 2016-04-22 20:20:31]

The CQ bit was checked by jmadill@chromium.org

[commit-bot: I haz the power, 2016-04-22 20:20:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1912153003/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1912153003/1

[commit-bot: I haz the power, 2016-04-22 22:03:31]

Description was changed from

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
==========

to

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
==========

[commit-bot: I haz the power, 2016-04-22 22:03:32]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-04-22 22:05:16]

Description was changed from

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987
==========

to

==========
Pin dEQP to ANGLE's DEPS file.

This will make rolling a new dEQP version in ANGLE testable on the
bots.

BUG=605987

Committed: https://crrev.com/60783171abc3fc122aeefd9a3705193e9191ccd9
Cr-Commit-Position: refs/heads/master@{#389262}
==========

[commit-bot: I haz the power, 2016-04-22 22:05:17]

Patchset 1 (id:??) landed as
https://crrev.com/60783171abc3fc122aeefd9a3705193e9191ccd9
Cr-Commit-Position: refs/heads/master@{#389262}

[kjellander_chromium, 2016-04-23 19:59:54]

kjellander@chromium.org changed reviewers:
+ kjellander@chromium.org

[kjellander_chromium, 2016-04-23 19:59:55]

agable/iannucci: I thought the "From" feature in DEPS files was deprecated
(WebRTC were using it in a while back and were recommended not to). Or has
things changed.

I'm wondering since this broke our auto-roll script. It's easy to update it ot
handle this, but I wanted to check first.

[iannucci, 2016-04-23 20:02:17]

iannucci@chromium.org changed reviewers:
+ iannucci@chromium.org

[iannucci, 2016-04-23 20:02:18]

Hm... I'm pretty sure that's deprecated. AFAIK, it doesn't work with git anyway?

[Jamie Madill, 2016-04-23 21:22:54]

The git diff step is also having problems on the ANGLE CQ, not sure if this is
related. (http://crbug.com/606150)

Is there a supported way of pulling in a dep from another module?

[iannucci1, 2016-04-23 21:43:30]

On 2016/04/23 21:22:54, Jamie Madill wrote:
> The git diff step is also having problems on the ANGLE CQ, not sure if this is
> related. (http://crbug.com/606150)
> 
> Is there a supported way of pulling in a dep from another module?

Not sure what you mean by 'module', but you can get the gclient dependencies of
some other repo by including that entire repo and then using `recursedeps`, ex:
https://chromium.googlesource.com/infra/infra/+/master/DEPS#159 . Note, however,
that due to the weirdness of gclient, all solution paths are relative to the
.gclient file. This means that if you have some projects with dependencies:

  foo: {foo/dep1, foo/dep2}
  bar: {bar/foo}
  recursedeps = ["bar/foo"]

You'll end up with:
  foo/dep1
  foo/dep2
  bar/foo

Which probably means that 'foo' will not be able to find its dependencies in the
correct place (e.g. relative to its own root). This essentially means that every
gclient project has opinions about where its toplevel folder is located relative
to the .gclient file. To make the above example work, you would want to:

  foo: {foo/dep1, foo/dep2}
  bar: {foo}
  recursedeps = ["foo"]

And commit a symlink `./bar/foo -> ./foo`. You'd end up with:
  foo/
  foo/dep1
  foo/dep2
  bar/
  bar/foo -> foo/

Hopefully that helps?

[Jamie Madill, 2016-04-24 01:29:53]

On 2016/04/23 21:43:30, iannucci1 wrote:
> On 2016/04/23 21:22:54, Jamie Madill wrote:
> > The git diff step is also having problems on the ANGLE CQ, not sure if this
is
> > related. (http://crbug.com/606150)
> > 
> > Is there a supported way of pulling in a dep from another module?
> 
> Not sure what you mean by 'module', but you can get the gclient dependencies
of
> some other repo by including that entire repo and then using `recursedeps`,
ex:
> https://chromium.googlesource.com/infra/infra/+/master/DEPS#159 . Note,
however,
> that due to the weirdness of gclient, all solution paths are relative to the
> .gclient file. This means that if you have some projects with dependencies:
> 
>   foo: {foo/dep1, foo/dep2}
>   bar: {bar/foo}
>   recursedeps = ["bar/foo"]
> 
> You'll end up with:
>   foo/dep1
>   foo/dep2
>   bar/foo
> 
> Which probably means that 'foo' will not be able to find its dependencies in
the
> correct place (e.g. relative to its own root). This essentially means that
every
> gclient project has opinions about where its toplevel folder is located
relative
> to the .gclient file. To make the above example work, you would want to:
> 
>   foo: {foo/dep1, foo/dep2}
>   bar: {foo}
>   recursedeps = ["foo"]
> 
> And commit a symlink `./bar/foo -> ./foo`. You'd end up with:
>   foo/
>   foo/dep1
>   foo/dep2
>   bar/
>   bar/foo -> foo/
> 
> Hopefully that helps?

It's definitely very helpful, yes. I'll have to puzzle together how to fix my
situation. Symlinks work as expected on all the relevant platforms? We do have
different paths for ANGLE's DEPS and for Chromium's DEPS.

[Jamie Madill, 2016-04-24 01:30:17]

On 2016/04/24 01:29:53, Jamie Madill wrote:
> On 2016/04/23 21:43:30, iannucci1 wrote:
> > On 2016/04/23 21:22:54, Jamie Madill wrote:
> > > The git diff step is also having problems on the ANGLE CQ, not sure if
this
> is
> > > related. (http://crbug.com/606150)
> > > 
> > > Is there a supported way of pulling in a dep from another module?
> > 
> > Not sure what you mean by 'module', but you can get the gclient dependencies
> of
> > some other repo by including that entire repo and then using `recursedeps`,
> ex:
> > https://chromium.googlesource.com/infra/infra/+/master/DEPS#159 . Note,
> however,
> > that due to the weirdness of gclient, all solution paths are relative to the
> > .gclient file. This means that if you have some projects with dependencies:
> > 
> >   foo: {foo/dep1, foo/dep2}
> >   bar: {bar/foo}
> >   recursedeps = ["bar/foo"]
> > 
> > You'll end up with:
> >   foo/dep1
> >   foo/dep2
> >   bar/foo
> > 
> > Which probably means that 'foo' will not be able to find its dependencies in
> the
> > correct place (e.g. relative to its own root). This essentially means that
> every
> > gclient project has opinions about where its toplevel folder is located
> relative
> > to the .gclient file. To make the above example work, you would want to:
> > 
> >   foo: {foo/dep1, foo/dep2}
> >   bar: {foo}
> >   recursedeps = ["foo"]
> > 
> > And commit a symlink `./bar/foo -> ./foo`. You'd end up with:
> >   foo/
> >   foo/dep1
> >   foo/dep2
> >   bar/
> >   bar/foo -> foo/
> > 
> > Hopefully that helps?
> 
> It's definitely very helpful, yes. I'll have to puzzle together how to fix my
> situation. Symlinks work as expected on all the relevant platforms? We do have
> different paths for ANGLE's DEPS and for Chromium's DEPS.

Also I think that means we should probably revert this change..

[Jamie Madill, 2016-04-24 01:30:54]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/1914843003/ by jmadill@chromium.org.

The reason for reverting is: Using unsupported syntax, could possibly be
breaking the ANGLE CQ.

BUG=606150.