GitLab

The ICU-based converter and the ICU-based IDNA logic use different
datasets from ICU.  This allows external embedders that only need IDNA
to avoid depending on huge (over 2 MiB) character set conversion
dataset.

Change-Id: I9e61208051d9553ddf93ec517ea07a7e54410e5e
Reviewed-on: https://chromium-review.googlesource.com/1185215Reviewed-by: Chris Palmer <palmer@chromium.org>
Commit-Queue: Victor Vasiliev <vasilvv@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585989}