This bug hasn't been looked at for a long time. Does gclient diff still crash now?

Yes.

[0:00:08] index 74be2a92..e923aab6 100644
[0:00:08] --- a/openh264.rc
[0:00:08] +++ b/openh264.rc
[0:00:08] @@ -24,8 +24,8 @@ LANGUAGE LANG_ENGLISH, SUBLANG_ENGLISH_US
[0:00:08]  //
[0:00:08]  
[0:00:08]  VS_VERSION_INFO VERSIONINFO
[0:00:08] - FILEVERSION 1,6,0,0
[0:00:08] - PRODUCTVERSION 1,6,0,0
[0:00:08] + FILEVERSION 1,7,0,1703
[0:00:08] + PRODUCTVERSION 1,7,0,1703
[0:00:08]   FILEFLAGSMASK 0x3fL
[0:00:08]  #ifdef _DEBUG
[0:00:08]   FILEFLAGS 0x1L
[0:00:08] @@ -43,12 +43,12 @@ BEGIN
[0:00:08]              VALUE "Comments", "Cisco OpenH264 codec"
[0:00:08]              VALUE "CompanyName", "Cisco Systems Inc."
[0:00:08]              VALUE "FileDescription", "Cisco OpenH264 codec"
[0:00:08] -            VALUE "FileVersion", "1.6.0.0"
[0:00:08] +            VALUE "FileVersion", "1.7.0.1703"
[0:00:08]              VALUE "InternalName", "openh264.dll"
----------------------------------------
Traceback (most recent call last):
  File "/usr/local/google/home/tnagel/depot_tools/gclient.py", line 2679, in <module>
    sys.exit(main(sys.argv[1:]))
  File "/usr/local/google/home/tnagel/depot_tools/gclient.py", line 2665, in main
    return dispatcher.execute(OptionParser(), argv)
  File "/ssd/tnagel/depot_tools/subcommand.py", line 252, in execute
    return command(parser, args[1:])
  File "/usr/local/google/home/tnagel/depot_tools/gclient.py", line 2463, in CMDdiff
    return client.RunOnDeps('diff', args)
  File "/usr/local/google/home/tnagel/depot_tools/gclient.py", line 1497, in RunOnDeps
    work_queue.flush(revision_overrides, command, args, options=self._options)
  File "/ssd/tnagel/depot_tools/gclient_utils.py", line 1059, in run
    self.item.run(*self.args, **self.kwargs)
  File "/usr/local/google/home/tnagel/depot_tools/gclient.py", line 885, in run
    file_list)
  File "/ssd/tnagel/depot_tools/gclient_scm.py", line 156, in RunCommand
    return getattr(self, command)(options, args, file_list)
  File "/ssd/tnagel/depot_tools/gclient_scm.py", line 278, in diff
    self._Run(['diff'] + merge_base, options)
  File "/ssd/tnagel/depot_tools/gclient_scm.py", line 1221, in _Run
    gclient_utils.CheckCallAndFilterAndHeader(cmd, env=env, **kwargs)
  File "/ssd/tnagel/depot_tools/gclient_utils.py", line 314, in CheckCallAndFilterAndHeader
    return CheckCallAndFilter(args, **kwargs)
  File "/ssd/tnagel/depot_tools/gclient_utils.py", line 534, in CheckCallAndFilter
    filter_fn(in_line)
  File "/ssd/tnagel/depot_tools/gclient_utils.py", line 309, in filter_msg
    filter_fn(line)
  File "/ssd/tnagel/depot_tools/gclient_utils.py", line 589, in __call__
    esc = line.find(unichr(033))
UnicodeDecodeError: 'utf8' codec can't decode byte 0xa9 in position 38: invalid start byte

Adding Infra>SDK because gclient is an "SDK" tool. It's probably a silly bug in gclient which expects unicode from git output or something like that. Given the low priority, I'm marking it available for now. We'll revisit it later if it becomes a bigger problem.

The invalid characters are showing up because it's diffing the entirety of fontconfig. It's doing that because "gclient diff" is doing the wrong thing.

Here's a fix for that: https://chromium-review.googlesource.com/#/c/chromium/tools/depot_tools/+/914717

There could still be a fix specifically for handling unicode characters in diffs better (although I think it does handle that fine from my testing; maybe it's having trouble with unicode filenames?) but this will resolve the major case.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/tools/depot_tools/+/1853f66efd16a0e0e6ad14cc75f2287a7fa19521

commit 1853f66efd16a0e0e6ad14cc75f2287a7fa19521
Author: Aaron Gable <agable@chromium.org>
Date: Tue Feb 13 00:40:17 2018

gclient: diff against hash in DEPS

When someone runs "gclient diff", they expect to be diffing their
local state against the state of a clean gclient checkout.

Up until now, "gclient diff" has been doing... not that. It's been
assuming that the sha1 in DEPS is approximately the head of the
remote's master branch, finding the merge-base between the local
content and that remote master, and diffing against that merge-base.
This makes basically no sense in the context of gclient. With
gclient, we already know exactly what should be checked out: it's
whatever is in DEPS.

This CL updates the behavior of "gclient diff" to do the right thing.

Bug: 731701
Change-Id: I23a43cbc0d6bc7772495097ff615d4d070f5ad19
Reviewed-on: https://chromium-review.googlesource.com/914717
Reviewed-by: Andrii Shyshkalov <tandrii@chromium.org>
Commit-Queue: Aaron Gable <agable@chromium.org>

[modify] https://crrev.com/1853f66efd16a0e0e6ad14cc75f2287a7fa19521/gclient_scm.py

Now that the immediate issue is solved, dropping this back into the queue of things we'd like to fix but I don't have time to take on right now :)