Summary:
Syncback tasks that move messages to a different folder used to leave
those messages without a uid because it was unknown at the moment, and
it would only be discovered later on in the sync loop.
This had the unwanted effect of not allowing you to perform more than 1
syncback action on the same thread back to back (e.g. undoing an archive, or
sending and archiving immediately)
This commit makes it so that `SetThreadFolderAndLabels` and
`MoveThreadToFolder` runs a new sync task to fetch the new uids for the
moved messages.
We do this via a new sync task, `FetchNewMessagesInFolder` which /only/
fetches new messages in a folder (no fetching old messages or attribute
updates). This is a first step to cleaning up the gigantic
`FetchMessagesInFolder` task into smaller parts-- but that will come in
a separate diff. For now we want to fix the immediate problem.
See D3788 and D3789 for more details
Test Plan:
manually move threads around, undo moving threads, reply on a thread
and immediately archive
Reviewers: khamidou, mark, spang, halla, evan
Reviewed By: evan
Differential Revision: https://phab.nylas.com/D3798
Summary:
updateMessagesFromThread requires `messages` to be an array only when
recompute is falsy
Test Plan: manual
Reviewers: halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D3797
Summary:
If there were any uidvalidity changes after we've completely synced a folder, we
would completely ignore them and not attempt to sync the folder.
Also, we weren't saving the latest uidvalidity from the box to the folder
syncState as soon as we recovered, we only saved it until after fetching
messages. This meant that if the operation was interrupted before updating
syncState.uidvalidity, we would always think that we were in a state of uid invalidity
Recovering from uidvalidity was also broken because we weren't resetting the
fetched ranges, and unnecessarily setting the folderId to null, which meant that
we would never restore the uids of messages we had already fetched.
Test Plan: manual
Reviewers: evan, khamidou, mark, halla
Reviewed By: mark
Differential Revision: https://phab.nylas.com/D3796
Summary:
Given that we perform syncback actions optimistically, i.e. we save the changes to the database before they are synced, some messages might not have an IMAP UID if they haven't been synced in the loop.
When we encounter messages without uid, we could just skip them, but this would mean that we end up moving only a subset of the messages we intended to move, or not move any at all, while thinking that we /did/ move everything, which might cause the thread to bounceback later on when we sync the messages that were missing.
The permanent fix is for syncback actions to not succeed until their changes have been synced, which is coming up in a separate diff
Test Plan: manual
Reviewers: evan, spang, khamidou, mark
Reviewed By: mark
Subscribers: mark
Differential Revision: https://phab.nylas.com/D3789
Summary:
In JavaScript, null <= null is truthy. So if you set fetchedmax but not
fetchedmin and minUID, and then highestmodseq doesn't increment, we will
incorrectly skip updates on All Mail after the first batch of Gmail
inbox UID downloads, only making further progress on sync if the user
receives a new mail.
This patch tightens isSyncComplete() to require that all variables used
in the math comparisons are set.
Test Plan: manual 😢
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3787
Summary:
Previously we just assumed that we were in the "charging" state and wouldn't
update that until the state changed. This would cause us to throttle even if
the app was opened while plugged in. Now we don't do that.
Test Plan: Run locally, verify that we no longer throttle
Reviewers: spang, evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3782
Summary:
Previously, when updating message attributes during folder sync, we
would fetch local db messages only by folderImapUID. This was incorrect,
because messages in different folders could have the same uid, so our
query would return messages from other folders, causing us to
incorrectly update their attributes.
This commit ensures we fetch local messages in the current folder
Test Plan: manual
Reviewers: evan, halla, mark, spang
Reviewed By: spang
Differential Revision: https://phab.nylas.com/D3777
We now correctly `await` for the `setFolders` and `setLabels`
operations, and make sure that the thread is always saved.
Apparently the `thread.id` check wasn't working as expected and we were skipping a
thread save, which only became apparent when we awaited for `setFolders`
and `setLabels`. Now, we always save the thread (like we did before this
method got updated)
Summary:
When I tested if we could use labels in place of the flag, I had the
Drafts folder selected, not All Mail, and the label did not show up.
Turns out that if you have All Mail selected, all draft messages will be
correctly tagged with the Draft label.
With this patch, we should correctly sync isDraft on all supported
providers and suppress these messages from the app until if or when we
decide to implement draft sync.
Thanks Brandon Long from the Gmail team for the help in getting this right.
Test Plan: manual
Reviewers: juan, evan
Reviewed By: juan, evan
Differential Revision: https://phab.nylas.com/D3775
Summary:
Otherwise accounts with `Sent` and an `Outbox` will both get the `sent`
role, causing inconsistencies in sync
Addresses T7682
Test Plan: manual
Reviewers: khamidou, spang
Reviewed By: khamidou, spang
Differential Revision: https://phab.nylas.com/D3773
Summary:
Now that message processing is serialized, this is unnecessary. I
was debugging something else and watched the code step through the
exact same db query twice in a row.
Test Plan: manual
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3764
Summary:
It turns out we don't ever use the /account API endpoint.
The GET /accounts endpoint was designed to query all accounts connected to
a system for old K2, but we don't use that anymore. The environment
variable protecting the endpoint isn't set anywhere.
We used to `GET /account` from the N1 AccountStore to attempt to refresh
the health of the accounts. The accompianing diff to this one makes that
obsolete. We never need to query for the account health since the sync
loop pushes it to us through `Actions.updateAccount`.
We also never `DELETE /account` because our local-sync runs a function
called `ensureK2Consistency` that compares its DB against the Nylas Mail
`AccountStore`.
This file has always been a huge source of confusion to me and a massive
red herring for anyone trying to understand how the account system work.
The accompyaning diff has more comments explaining the existing system
Depends on https://phab.nylas.com/D3770
Test Plan:
Manually boot N1. Ensure existing account works. Add a new account. Remove
an account. Open the developer tools and check that all the tabs still
work.
Lots of grepping through the code base.
Reviewers: halla, mark, juan, khamidou
Reviewed By: juan, khamidou
Differential Revision: https://phab.nylas.com/D3769
Summary:
This will solve T7579 when saving messages to the sent folder. I
attempted to clean up the references code but decided it was better left for a
new diff, so added a bunch of TODO's in this diff
Test Plan: manual
Reviewers: halla, spang, evan
Reviewed By: spang, evan
Differential Revision: https://phab.nylas.com/D3766
Summary:
We've been syncing drafts messages but not the drafts flag in K2, making
them appear in Edgehill as regular old messages.
This commit makes K2 sync the drafts flag, and also correctly label
folders called "Drafts" with the 'drafts' role.
Because 2-way syncing of drafts is very complex and error-prone since
you need to add new drafts and delete the old ones on every update, and
we reaally don't want to do things like create multiplying draft copies
or accidentally lose a draft someone started composing elsewhere, we
simply exclude messages marked as drafts from being serialized to
Edgehill through the delta stream for now. This removes the confusing
behaviour and also sets a better stage for completing drafts sync later.
Eventually we will also want to add functionality to allow users to
select their drafts folder, but for now this code does the right thing
in many more cases.
While investigating this behaviour, I also discovered a bug we've never
seen before where Gmail isn't applying the \Draft flag to draft
messages, no matter which folder we fetch them from. :-/ This is very
unfortunate and there's no way for us to work around it other than to
fetch messages in the Drafts folder and manually apply the flag locally,
since "drafts" is not a label in Gmail, only another IMAP folder. Brandon
Long from the Gmail team says that this is because they've had
problems with clients which sync drafts, so the Gmail web client and
mobile apps do not set the \Draft flag on drafts. (I don't get how this
solves their problem, but okay.) Let's solve the issue on Gmail if it
comes up by user demand—should be relatively straightforward to
implement, but it adds sync work & complexity.
Fixes T7593
Test Plan: manual
Reviewers: halla, juan
Reviewed By: juan
Maniphest Tasks: T7593
Differential Revision: https://phab.nylas.com/D3749
Summary:
We don't want to run message processing full tilt when a user isn't plugged in.
This diff adds some detection logic that causes message processing to be
throttled/unthrottled when a user unplugs/plugs in their computer.
Test Plan: Run locally unplugged and plugged in, verify that CPU use goes up/down
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3759
Summary:
See title
Depends on D3744
Test Plan: tested locally
Reviewers: spang, evan, juan
Reviewed By: spang, evan, juan
Differential Revision: https://phab.nylas.com/D3745
Summary:
I thought it was gonna be OK that we kept all HTML parts in a multipart/alternative
MIME structure because the world is a sane place and nobody would ever put more
than one HTML part in a multipart/alternative structure.
I was wrong.
We have found extraterrestrial life^W^WI mean emails which contain duplicate,
exactly the same MIME parts within a multipart/alternative MIME structure: two
text/plain parts and two text/html parts. This is likely due to a broken MIME
implementation, or perhaps a bug in someone's email script. So, we should
only keep one text/html MIME part if there are multiple.
Test Plan:
manual for now---added this to my mail parsing regression test list
for implementation once we unify the DBs and have a roughly stable code
structure
Reviewers: halla, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3750
Summary:
Rather than having a strict model where we don't decode the message if we
don't specifically recognize the CTE, treat any CTEs we don't recognize as
having no encoding. There are several CTE strings that could mean this (e.g.
7bit, 7BITS, 8-bit, binary, NONE, utf8), and we don't want to check for them
all. Additionally, if there is a CTE we don't support, the user will likely
see rendering issues and contact support. This will allow us to obtain more
concrete information about these messages.
Test Plan: manual
Reviewers: spang
Reviewed By: spang
Differential Revision: https://phab.nylas.com/D3748
Summary:
Given that we were marking the account as errored if we've encountered
enough RetryableErrors, we would show the red box to the user when in
fact the problem was the user was offline, causing confusion
If the user is offline, we will constantly get RetryableErrors in the
sync loop, and we can't mark the account as errored in that case.
Test Plan: manual
Reviewers: evan
Reviewed By: evan
Differential Revision: https://phab.nylas.com/D3752
Summary:
This commit ensures that we handle transient errors correctly when refreshing
tokens
Test Plan: manual
Reviewers: khamidou, evan
Reviewed By: evan
Differential Revision: https://phab.nylas.com/D3740
Summary:
In the sync worker:
- Move the backoff logic inside `scheduleNextSync`, where all logic to schedule the next sync loop now lives
- If we've retried a RetryableError a bunch of times, show the error to the user, otherwise the user might think the app is not working for no reason
- Clean up logging
In the message processor:
- Report message processing errors to sentry!
Sync Process Manager:
- Listen to new `Actions.debugSync` to show the Activity Window and open dev tools
Test Plan: manual
Reviewers: khamidou, evan
Reviewed By: khamidou, evan
Differential Revision: https://phab.nylas.com/D3736
Summary:
This commit makes it so our syncback tasks send as few imap commands as possible by passing a set of UIDs whenever possible. Previously, we would send 1 command per message, with a single UID, which was very wasteful given that we can pass a set of UIDs. This is especially helpful for operating on threads with a large number of messages.
Syncback actions will now group all messages in a thread by the folder they belong to, and issue a single operation on the folder box. When removing all labels from a thread (setting labels to []), we need to issue a command of the form `box.delLabels(uids, labels)`, so we also group messages by set their set of labels to issue as few commands as possible.
This commit only batches imap commands, but we can still batch syncback actions themselves, which can be implemented in a separate patch.
Test Plan: manual
Reviewers: evan, mark, spang
Reviewed By: spang
Subscribers: halla, mg
Differential Revision: https://phab.nylas.com/D3719
Summary: Double-firing protection since the DatabaseStore can now fire this
Test Plan: manual
Reviewers: juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3730
Summary:
We weren't, which meant that us sending with multi-send or generic IMAP
broke threading. :(
Test Plan: manual
Reviewers: juan, evan
Reviewed By: juan, evan
Differential Revision: https://phab.nylas.com/D3718
Summary:
Before this commit, if folder sync was complete, and the account didn't support CONDSTORE (e.g. Office365, Yahoo), we would only check for attribute updates every 10 minutes.
This commit makes it so we always check for attribute updates if the server doesn't support CONDSTORE
So for example, when marking a thread as read, we would perform the optimistic update in N1, queue the syncback task which would succeed, but the thread in k2s db would never get updated and become stale, with an unreadCount > 0. If we emitted a delta for that thread during the window of time where we ignored attribute updates, it would be set as unread again in N1, even though all of its messages were read.
This still doesn't guarantee that it wont happen (we could still get a delta for the thread before we actually fetch the attribute updates from IMAP), but before this commit it was sure to happen. This should be properly fixed with the sync scheduler refactor
Test Plan: manual
Reviewers: evan, mark, spang
Reviewed By: mark, spang
Differential Revision: https://phab.nylas.com/D3714
Summary:
Previously, we were not pripritizing archive sync when getting folders to sync, causing it to be synced almost last. I believe this was causing the issues regarding archived items coming back, because we would optimistically archive in N1, but the changes wouldn't be reflected in K2's database until we synced the archive, causing the data to become out of sync. If for whatever reason we got a delta for any of those messages before the archive was synced, they would pop back in the inbox because in k2, they were still in the inbox. This was exacerbated by the fact that all syncback tasks would interrupt the loop, so we would reach the archive until very late, making this scenario way more likely.
This still wont guarantee that it wont happen, because we dont do /any/ optimistic updates in K2, so we could still get deltas before we actually sync the folder, but makes the scenario way less likely. This should be properly fixed with the sync scheduler refactor
Test Plan: manual
Reviewers: spang, evan, mark
Reviewed By: mark
Differential Revision: https://phab.nylas.com/D3716
Summary:
All we do is use the SEARCH X-GM-RAW IMAP extension to find the UIDs
to prioritize at the beginning of initial sync, and download these UIDs
until there are none left. Then we continue downloading All Mail as
usual.
Because of the way we batch via ranges, the most expedient way to
implement this means that all prioritized emails will end up being
downloaded twice (the second time we'll detect that the message exists
and do nothing).
This seems like a worthwhile tradeoff for quick appearance of the
messages in a user's inbox.
Test Plan: manual
Reviewers: evan, juan
Reviewed By: evan, juan
Differential Revision: https://phab.nylas.com/D3706
Summary:
Don't show the attachment icon on threads that only have inline
images. We do this by assuming that inline images have a contentID,
and regular attachments do not. Also updates the way we send
attachments in order to adhere to this standard.
Test Plan: tested manually
Reviewers: spang
Reviewed By: spang
Differential Revision: https://phab.nylas.com/D3696
Summary:
When syncing folders, we check if the folder needs syncing by checking if it has any new messages via the STATUS command (STATUS returns uidnext, highestmodseq among others, and is cheaper than SELECT)
However, we can't issue a STATUS on a box that is already selected. Previously, if the box was already selected, we would just return it, but this was incorrect because we wouldn't get the latest box values (e.g. uidnext), causing us to think that there were no updates available, and skip syncing folders that actually needed to be synced.
Now, if the box is already selected when getting the status, we have to re select it to refresh the latest values
Test Plan: manual
Reviewers: evan, khamidou, spang
Reviewed By: spang
Differential Revision: https://phab.nylas.com/D3697
Summary:
This commit also lowers the batch size of messages to fetch on folder sync down to 30. This is in order to prevent sync from getting stuck if we queue too many syncback tasks-- given that we only update the range of fetched uids after we've actually fetched and processed messages, if the batch size is too big and we interrupt too often, we might end up never advancing the range and re fetching the same messages over and over.
This also makes the sync loop run faster through all folders in general.
Depends on D3689 to make sure that the batch size actually reflects a message count, i.e. to ensure that we are making /visible/ progress.
Test Plan: manual
Reviewers: spang, khamidou, evan
Reviewed By: evan
Maniphest Tasks: T7477
Differential Revision: https://phab.nylas.com/D3692
Summary:
Because we optimistically fetch UIDs by expanding a range without looking
at the actual UIDs in the inbox and the actual space of UIDs with messages
attached may be sparse due to message moves, we need to track how many
messages we actually download during a range expansion and continue
expanding the range if we haven't downloaded enough messages.
If we reach a large gap where we download no messages at all during a batch, we
pause and check the actual UID list for the folder for the next UID to
download, as otherwise we may spin indefinitely fetching UIDs that don't exist.
(Example: my "Deleted Items" folder had about 300k worth of empty UIDs between
a very small UID and a very large UID. With the new system, this registers as a
completed sync within a single iteration as soon as sync hits the gap.)
Test Plan: manual
Reviewers: juan, evan
Reviewed By: juan, evan
Differential Revision: https://phab.nylas.com/D3689
Summary:
This patch changes the sync worker to back off exponentially when there is an issue syncing an account. This has two goals:
- first, it's a bit dangerous to retry immediately. We don't want hundreds of thousands of machines trying to refresh tokens unsuccessfully because our service is struggling.
- second, it's nicer on the CPU to wait a bit between retries.
Currently, we sleep for at most 2 minutes, with some random jitter added.
Test Plan: Tested manually, stared at the code a long time.
Reviewers: evan, juan
Reviewed By: evan, juan
Differential Revision: https://phab.nylas.com/D3684
