Yep, here's the diff of new initializers:

# debug.cpp std::ios_base::Init::Init()@plt
# debug.cpp std::__ioinit [#includes <iostream>, use <ostream> instead]
# debug.cpp __cxa_atexit@plt [registers a dtor to run at exit]
# debug.cpp __dso_handle

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8cc767cb0d1c65af6d5a36858c9633c929b7e4cd

commit 8cc767cb0d1c65af6d5a36858c9633c929b7e4cd
Author: findit-for-me <findit-for-me@appspot.gserviceaccount.com>
Date: Tue May 16 23:14:54 2017

Revert of Enable DCHECKs by default in non-official builds. (patchset #3 id:40001 of https://codereview.chromium.org/2886803002/ )

Reason for revert:

Findit (https://goo.gl/kROfz5) identified CL at revision 472195 as the
culprit for failures in the build cycles as shown on:
https://findit-for-me.appspot.com/waterfall/culprit?key=ag9zfmZpbmRpdC1mb3ItbWVyRAsSDVdmU3VzcGVjdGVkQ0wiMWNocm9taXVtLzg3NWEyZmE0M2NhYWIwZWU4MmVmZTY4ZmI1YWRiMzU3MzkyNDM0YWQM

Original issue's description:
> Enable DCHECKs by default in non-official builds.
>
> BUG= 722827 
> TBR=sdefresne@chromium.org for ios/
>
> Review-Url: https://codereview.chromium.org/2886803002
> Cr-Commit-Position: refs/heads/master@{#472195}
> Committed: https://chromium.googlesource.com/chromium/src/+/875a2fa43caab0ee82efe68fb5adb357392434ad

TBR=brettw@chromium.org,thakis@chromium.org,sdefresne@chromium.org,gab@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 722827 , 723090 , 723049 

Review-Url: https://codereview.chromium.org/2892473002
Cr-Commit-Position: refs/heads/master@{#472242}

[modify] https://crrev.com/8cc767cb0d1c65af6d5a36858c9633c929b7e4cd/build/config/dcheck_always_on.gni
[modify] https://crrev.com/8cc767cb0d1c65af6d5a36858c9633c929b7e4cd/ios/web/web_state/ui/wk_web_view_configuration_provider.mm

debug.cpp = third_party/angle/src/common/debug.cpp ?

Thinking that because it's the only debug.cpp in codebase which includes iostream as the logs indicate is at fault.

use <ostream> instead then?

Since I couldn't lower the expectation count to 6, this is actually the second time SI count has increased. Sigh.

Removing from sheriff queue since this isn't currently blocking the three

I think you can resume work on this, the ANGLE bug is fixed.

I think that was the only blocker, thanks!