Observation: crash/51105c7878000000
CHROMEOS_RELEASE_VERSION: 9592.42.0
CHROME VERSION: 60.0.3112.52 beta

There's another bug for the leaf function open (crbug.com/533262), but it's old (since M-52) and doesn't follow the same stack trace.

Thread 0 (id: 24541) CRASHED [SIGSEGV @ 0x00000138 ] MAGIC SIGNATURE THREAD
0x00005b8bbe9d6e44	(chrome + 0x04b25e44 )	extensions::Extension::GetManifestData(std::string const&) const
0x00005b8bbea0011c	(chrome + 0x04b4f11c )	extensions::IconsInfo::GetIcons(extensions::Extension const*)
0x00005b8bbedab540	(chrome + 0x04efa540 )	extensions::ChromeAppIcon::Reload()

It seems possible that we're passing a null extension:
void ChromeAppIcon::Reload() {
  const Extension* extension = GetExtension();
  icon_ = base::MakeUnique<IconImage>(
      browser_context_, extension, IconsInfo::GetIcons(extension),
      resource_size_in_dip_, util::GetDefaultAppIcon(), this);
  UpdateIcon();
}

// static
const ExtensionIconSet& IconsInfo::GetIcons(const Extension* extension) {
  IconsInfo* info = static_cast<IconsInfo*>(
      extension->GetManifestData(keys::kIcons));
  return info ? info->icons : g_empty_icon_set.Get();
}

I believe that https://codereview.chromium.org/2900783003 fixed this at ToT by handling the NULL extension case properly, but that code is newer than M-60.

CCing khmel to see if that sounds reasonable.

CCing bhthompson about impact--
I don't know how to evaluate if this is worth addressing for 60.

Here are crash statistics for the flow through GetIcons:
https://crash.corp.google.com/browse?q=OMIT%20RECORD%20IF%20SUM(CrashedStackTrace.StackFrame.FunctionName%3D%27extensions%3A%3AIconsInfo%3A%3AGetIcons(extensions%3A%3AExtension%20const*)%27)%20%3D%200&ignore_case=false&enable_rewrite=true&omit_field_name=CrashedStackTrace.StackFrame.FunctionName&omit_field_value=extensions%3A%3AIconsInfo%3A%3AGetIcons(extensions%3A%3AExtension%20const*)&omit_field_opt=%3D#samplereports