All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=873712

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=28ba97f8571afbb8f518e562afc8b9b186b9f42306244df7c5e8c5ea855caedd


Bot(s) for this bug's original alert(s):

Android Builder Perf

Assigning to jperaza@chromium.org because this is the only CL in range:
Add CrashReportMimeWriter to MIME encode crashpad minidumps

CrashReportMimeWriter provides a method to call into native code to
rewrite the minidumps as MIME messages. This CL also links that native
code into webview and clank, but makes no functional changes.

Bug: crashpad:30
Change-Id: Ief41d366614a833db3b643f46a0297c52fea495f
Reviewed-on: https://chromium-review.googlesource.com/1115702
Commit-Queue: Joshua Peraza <jperaza@chromium.org>
Reviewed-by: Ilya Sherman <isherman@chromium.org>
Reviewed-by: Peter Wen <wnwen@chromium.org>
Reviewed-by: agrieve <agrieve@chromium.org>
Reviewed-by: Mark Mentovai <mark@chromium.org>
Reviewed-by: Bo <boliu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#581918}

Can see the symbol diff here:
https://logs.chromium.org/logs/chromium/buildbucket/cr-buildbucket.appspot.com/8938634272010227104/+/steps/Show_Supersize_Diff/0/logs/s_____Show_Diff____/0

(android_binary_size bot from the "Show Experimental" Gerrit link).

It's all just new code from crashpad. Josh - can you just sanity check the list of symbols to see if they look expected? There are certainly some base-type symbol util type things in there, but guessing crashpad cannot depend on base?

Yep, this looks expected to me. Crashpad does depend on base (it uses a miniaturized version of base when building externally). It looks like all the new base symbols are for uses of the templatized base::ScopedGeneric.

Ah, I gotcha. Looks like most of the base-like operations have the side-effect of additional logging, which I suppose is important for crash-related code.

Thanks for having a look!