GitLab

The comment specified the purpose and possible values of the --password-store
switch. It was outdated in both respects. Since the proper implementation of
os_crypt on Linux earlier this year, the storage was used not just for
passwords, but also for os_crypt's encryption key.

This is even more problematic because the comment is picked up to generate
http://peter.sh/experiments/chromium-command-line-switches/.

This CL updates the description based on the relevant code in
components/os_crypt/key_storage_util_linux.cc.

The CL does not change the name of the flag on purpose: it is likely contained
in a lot of testing and developers' set-ups, and changing it could have
unwanted side-effects of testing data suddenly piling up in people's
Keyrings/KWallets. Once the alternative backends for Linux will stop being
used for storing passwords (https://crbug.com/571003), and only for the single
encryption key of os_crypt, it will be much safer to rename the switch.

BUG=673121

Review-Url: https://codereview.chromium.org/2563253002
Cr-Commit-Position: refs/heads/master@{#437880}