Mailspring/internal_packages/worker-sync/lib/nylas-sync-worker-pool.coffee
Evan Morikawa 395bd80da2 fix(draft-list): draft list removes draft when the message sends
Summary:
This started when I noticed that drafts weren't dissapearing from the
draft list after send. This was a pretty big bug because if you ever
clicked on one again and tried to re-send it would throw a 400 error
saying the draft id doesn't exist.

This uncovered a few fundamental issues with the DB.

First of all, the reason the draft list wasn't updating was because the DB
trigger that happened when we got in a new message, was being ignored
since the diff contained no drafts (it's now a message).

The bigger issue was that if you had a draft with only a clientId, gave it
a serverId, and tried to call "save", the REPLACE INTO method would not
update the old object, but rather create a second duplicate. This is
because the `id` field was being used as the PRIMARY KEY, and in this
case, that `id` field changed! The fix was to change the PRIMARY KEY to be
the `cilent_id` instead of the `id` and use that as the REPLACE INTO
index.

We still need the `id` field; however, because all of our reads depend on
that field usually being the serverId

Fixes T3507

Test Plan: See new and updated tests

Reviewers: dillon, bengotow

Reviewed By: bengotow

Maniphest Tasks: T3507

Differential Revision: https://phab.nylas.com/D1992
2015-09-08 13:11:34 -07:00

128 lines
4.2 KiB
CoffeeScript

_ = require 'underscore'
{NylasAPI,
Actions,
AccountStore,
DatabaseStore,
DatabaseObjectRegistry} = require 'nylas-exports'
NylasLongConnection = require './nylas-long-connection'
NylasSyncWorker = require './nylas-sync-worker'
class NylasSyncWorkerPool
constructor: ->
@_workers = []
AccountStore.listen(@_onAccountsChanged, @)
@_onAccountsChanged()
_onAccountsChanged: ->
return if atom.inSpecMode()
accounts = AccountStore.items()
workers = _.map(accounts, @workerForAccount)
# Stop the workers that are not in the new workers list.
# These accounts are no longer in our database, so we shouldn't
# be listening.
old = _.without(@_workers, workers...)
worker.cleanup() for worker in old
@_workers = workers
workers: =>
@_workers
workerForAccount: (account) =>
worker = _.find @_workers, (c) -> c.account().id is account.id
return worker if worker
worker = new NylasSyncWorker(NylasAPI, account)
connection = worker.connection()
connection.onStateChange (state) ->
Actions.longPollStateChanged({accountId: account.id, state: state})
if state == NylasLongConnection.State.Connected
## TODO use OfflineStatusStore
Actions.longPollConnected()
else
## TODO use OfflineStatusStore
Actions.longPollOffline()
connection.onDeltas (deltas) =>
@_handleDeltas(deltas)
@_workers.push(worker)
worker.start()
worker
_cleanupAccountWorkers: ->
for worker in @_workers
worker.cleanup()
@_workers = []
_handleDeltas: (deltas) ->
Actions.longPollReceivedRawDeltas(deltas)
# Create a (non-enumerable) reference from the attributes which we carry forward
# back to their original deltas. This allows us to mark the deltas that the
# app ignores later in the process.
deltas.forEach (delta) ->
if delta.attributes
Object.defineProperty(delta.attributes, '_delta', { get: -> delta })
{create, modify, destroy} = @_clusterDeltas(deltas)
# Apply all the deltas to create objects. Gets promises for handling
# each type of model in the `create` hash, waits for them all to resolve.
create[type] = NylasAPI._handleModelResponse(_.values(dict)) for type, dict of create
Promise.props(create).then (created) =>
# Apply all the deltas to modify objects. Gets promises for handling
# each type of model in the `modify` hash, waits for them all to resolve.
modify[type] = NylasAPI._handleModelResponse(_.values(dict)) for type, dict of modify
Promise.props(modify).then (modified) =>
# Now that we've persisted creates/updates, fire an action
# that allows other parts of the app to update based on new models
# (notifications)
if _.flatten(_.values(created)).length > 0
Actions.didPassivelyReceiveNewModels(created)
# Apply all of the deletions
destroyPromises = destroy.map(@_handleDeltaDeletion)
Promise.settle(destroyPromises).then =>
Actions.longPollProcessedDeltas()
_clusterDeltas: (deltas) ->
# Group deltas by object type so we can mutate the cache efficiently.
# NOTE: This code must not just accumulate creates, modifies and destroys
# but also de-dupe them. We cannot call "persistModels(itemA, itemA, itemB)"
# or it will throw an exception - use the last received copy of each model
# we see.
create = {}
modify = {}
destroy = []
for delta in deltas
if delta.event is 'create'
create[delta.object] ||= {}
create[delta.object][delta.attributes.id] = delta.attributes
else if delta.event is 'modify'
modify[delta.object] ||= {}
modify[delta.object][delta.attributes.id] = delta.attributes
else if delta.event is 'delete'
destroy.push(delta)
{create, modify, destroy}
_handleDeltaDeletion: (delta) =>
klass = NylasAPI._apiObjectToClassMap[delta.object]
return unless klass
DatabaseStore.find(klass, delta.id).then (model) ->
return Promise.resolve() unless model
return DatabaseStore.unpersistModel(model)
pool = new NylasSyncWorkerPool()
window.NylasSyncWorkerPool = pool
module.exports = pool