Fix for building with ancient toolchain

Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none

Committed: https://crrev.com/0a23e5ae7723975148d7f7f3d427f435445e0c21
Cr-Commit-Position: refs/heads/master@{#440950}

[hugoh_UTC2, 2016-12-02 09:54:46]

Description was changed from

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none
==========

to

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none
==========

[hugoh_UTC2, 2016-12-02 09:54:46]

hugoh@opera.com changed reviewers:
+ pdr@chromium.org

[hugoh_UTC2, 2016-12-02 09:56:39]

Hi Philip! Here is a small fixup that we need after
https://codereview.chromium.org/2445413003 .

[pdr., 2016-12-02 21:51:53]

On 2016/12/02 at 09:56:39, hugoh wrote:
> Hi Philip! Here is a small fixup that we need after
https://codereview.chromium.org/2445413003 .

LGTM

How often do we break you like this? If it's a nontrivial amount, can we get our
infra team to setup a bot with your toolchain?

[Mostyn Bramley-Moore, 2016-12-06 00:04:24]

On 2016/12/02 21:51:53, pdr. wrote:
> How often do we break you like this? If it's a nontrivial amount, can we get
our
> infra team to setup a bot with your toolchain?

I estimate that such breakage (uClibc-related) occurs 2-5 times per month.  It's
normally trivial to fix but sometimes it's hard to motivate the effort on our
side to upstream the fixes, so we have been living with a bunch of local patches
to cope.  I think having an official bot testing this configuration would help a
lot.  

Setting up a bot would probably require some work on our side first though- we
would need to either put together a sysroot image that is capable of building
content_shell (our targets lack X11 and other desktop-oriented libraries), or
find another configuration (embedded content shell?) without the desktop-library
deps and produce a sysroot image for that.  We can't redistribute our existing
sysroot images without checking with legal first (they typically include code
from the SoC vendor, with non-free licenses).  Before going to the effort, I
would like some input from your infra team as to whether or not this would be
accepted.  

On a related note, we also see GCC-specific breakage now that all the official
linux bots use Clang (as far as I can tell).  That would be much easier to
arrange a semi-official bot for (the current Ubuntu LTS release would probably
suffice).


-Mostyn.

[pdr., 2016-12-06 20:00:56]

On 2016/12/06 at 00:04:24, mostynb wrote:
> On 2016/12/02 21:51:53, pdr. wrote:
> > How often do we break you like this? If it's a nontrivial amount, can we get
our
> > infra team to setup a bot with your toolchain?
> 
> I estimate that such breakage (uClibc-related) occurs 2-5 times per month. 
It's normally trivial to fix but sometimes it's hard to motivate the effort on
our side to upstream the fixes, so we have been living with a bunch of local
patches to cope.  I think having an official bot testing this configuration
would help a lot.  
> 
> Setting up a bot would probably require some work on our side first though- we
would need to either put together a sysroot image that is capable of building
content_shell (our targets lack X11 and other desktop-oriented libraries), or
find another configuration (embedded content shell?) without the desktop-library
deps and produce a sysroot image for that.  We can't redistribute our existing
sysroot images without checking with legal first (they typically include code
from the SoC vendor, with non-free licenses).  Before going to the effort, I
would like some input from your infra team as to whether or not this would be
accepted.  
> 
> On a related note, we also see GCC-specific breakage now that all the official
linux bots use Clang (as far as I can tell).  That would be much easier to
arrange a semi-official bot for (the current Ubuntu LTS release would probably
suffice).
> 
> 
> -Mostyn.

+cc Julie. I wanted to put this issue on your radar, though nothing needs to be
done at the moment. Is there anyone we could connect Mostyn with to help on this
kind of issue?

[hugoh_UTC2, 2016-12-29 09:28:52]

The CQ bit was checked by hugoh@opera.com

[commit-bot: I haz the power, 2016-12-29 09:29:02]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-12-29 10:37:40]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1483003732918250, "parent_rev":
"9f2de728baba0de50fa46d5190dcc947e55a73e0", "commit_rev":
"03d1c4698b273a3ff37aced02a6a7f34e6867760"}

[commit-bot: I haz the power, 2016-12-29 10:38:12]

Description was changed from

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none
==========

to

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none

Review-Url: https://codereview.chromium.org/2548783002
==========

[commit-bot: I haz the power, 2016-12-29 10:38:13]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2017-01-02 15:52:04]

Description was changed from

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none

Review-Url: https://codereview.chromium.org/2548783002
==========

to

==========
Fix for building with ancient toolchain

At Opera we test our SDK on an old MIPS device with a
peculiar toolchain. To be able to compile ImageDocument.o,
we need to call std::round(), not just round().

BUG=none

Committed: https://crrev.com/0a23e5ae7723975148d7f7f3d427f435445e0c21
Cr-Commit-Position: refs/heads/master@{#440950}
==========

[commit-bot: I haz the power, 2017-01-02 15:52:05]

Patchset 1 (id:??) landed as
https://crrev.com/0a23e5ae7723975148d7f7f3d427f435445e0c21
Cr-Commit-Position: refs/heads/master@{#440950}