GitLab

This reverts commit 232ad148.

Reason for revert: this patch was reverted in crrev.com/c/2043451 due to suspect of sync protocol violation. It was fixed in crrev.com/c/2061751.

Original change's description:
> Revert "Unify clearing of sync prefs"
> 
> This reverts commit 44d3aa82.
> 
> Reason for revert: suspect of running into sync protocol violations and the corresponding DCHECK failures.
> 
> Original change's description:
> > Unify clearing of sync prefs
> > 
> > SyncPrefs is a thin layer on top of preferences and hosts two groups
> > of preferences:
> > 1) Actual user-facing settings, such as type-selection, exposed via
> >    SyncUserSettings.
> > 2) Local "bookkeeping" sync metadata, such the last synced time or the
> >    client ID (cache GUID).
> > 
> > In addition, there are two cases that fall somewhere in the middle,
> > whose behavior is changed in this patch:
> > a) FirstSetupComplete: which roughly represents the user having
> >    consented to sync-the-feature (as opposed to transport-only upon
> >    sign-in without explicit user consent).
> > b) The encryption-bootstrap-token: which represent an explicit
> >    passphrase (usually custom passphrase) entered by the user, that
> >    allows decrypting the incoming sync changes and encrypt outgoing
> >    ones.
> > 
> > The last two preferences above fit group 1 better, so this patch stops
> > clearing them in SyncPrefs::ClearPreferences(), now renamed to
> > SyncPrefs::ClearLocalSyncTransportData().
> > 
> > With such cleanup, what used to be
> > ClearDirectoryConsistencyPreferences() is now merged into a single
> > clearing function, ClearLocalSyncTransportData(), and all calling sites
> > are unified by directly clearing prefs in ShutdownImpl(DISABLE_SYNC), as
> > opposed to individual calling sites.
> > 
> > This introduces -arguably desirable- behavioral changes because
> > codepaths like RESET_LOCAL_SYNC_DATA or STOP_SYNC_FOR_DISABLED_ACCOUNT
> > now actually clear all local metadata, which most notably includes
> > keystore keys.
> > 
> > Change-Id: I2c42f98c4e068c7e340580d0b78a5cd5b5c46171
> > Bug: 1046237
> > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2023649
> > Commit-Queue: Mikel Astiz <mastiz@chromium.org>
> > Reviewed-by: Marc Treib <treib@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#735841}
> 
> TBR=treib@chromium.org,mastiz@chromium.org
> 
> # Not skipping CQ checks because original CL landed > 1 day ago.
> 
> Bug: 1046237, 1048771
> Change-Id: I4ce9e941f12aa58b7b6b286e990e74a1e01dad69
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2043451
> Commit-Queue: Rushan Suleymanov <rushans@google.com>
> Reviewed-by: Mikel Astiz <mastiz@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#739374}

TBR=treib@chromium.org,mastiz@chromium.org,rushans@google.com

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 1046237, 1048771
Change-Id: Idfad70ab9b1c00f56dc8e78257fc1a75fc71466f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2066726Reviewed-by: Mikel Astiz <mastiz@chromium.org>
Commit-Queue: Rushan Suleymanov <rushans@google.com>
Cr-Commit-Position: refs/heads/master@{#743896}