Summary: See title. Part of T7681.
Test Plan: ran the specs
Reviewers: evan, mark, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4264
Summary:
Periodically ping client-sync's /health endpoint and store the latest
sync activity. If we get an ECONNREFUSED error, the worker window is
unavailable. Report the last known activity and restart the worker
window. Part of T7681.
Test Plan: manual, specs
Reviewers: evan, mark, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4263
Summary:
We still don't know exactly what scenarios cause us to get 'database is
locked' errors, but generically this means that we have multiple
processes trying to access the database at the same time. In an attempt
to handle this gracefully, this diff makes it so that we retry the
queries in these cases. Theoretically, the database should free up once
the other process is done using it, and the erroring process just needs
to wait its turn. We still throw an error after 5 retries, so if there's
a larger issue, we'll still be able to tell in Sentry.
Addresses T7992
Test Plan:
I opened a transaction in the worker window and then tried to
do the same in the main window. If I didn't release the transaction in
the worker window, the main window eventually errored. If I did release
the transaction, the main window continued creating its own transaction.
Reviewers: mark, juan, evan
Reviewed By: juan, evan
Differential Revision: https://phab.nylas.com/D4254
Summary:
We weren't doing that, so we would have accurate search index info for
threads when they were, e.g., moved to another folder.
Test Plan: Run locally
Reviewers: spang, evan, juan
Reviewed By: evan, juan
Differential Revision: https://phab.nylas.com/D4260
Summary:
Before this commit, it was impossible to remove inline images via the x
button
Test Plan: manual
Reviewers: mark, halla
Reviewed By: mark, halla
Differential Revision: https://phab.nylas.com/D4257
Summary:
I had a bit of downtime this morning so I decided to look into how to store Electron crash data. Electron, like Chromium, stores crash data in an arcane file format named minidump.
There's a bunch of services you can use to store files formatted in this format, like Mozilla's Socorro or zcbenz's own mini-breakpad-server but they're all pretty hard to install. Luckily, it turns out there's a Ruby gem to process minidump files and send them to Sentry! Seeing that, I whipped up a quick sinatra service and hosted it on Heroku – you can take a look at the source code here if you're curious: https://github.com/khamidou/electron-breakpad-sentry
This diff adds the client-side code we need to start sending crash reports.
Test Plan: Tested manually.
Reviewers: juan, evan
Reviewed By: juan, evan
Maniphest Tasks: T7926
Differential Revision: https://phab.nylas.com/D4259
Summary:
We'd like to get an idea of how long the client app is having to wait
when requesting a file. Waiting can cause things like inline image
attachments to appear very slow to load.
Test Plan: Run locally
Reviewers: evan, spang, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4255
Summary:
The original name seems like it's initiating the download, when really
it's just returning the data of an already in-progress/completed
download.
Test Plan: Manual, specs
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4251
Summary:
We need to download the files and then treat them as uploads. Rather
than using an actual Upload object, which would require another data
transfer, we create an object with all the necessary Upload-like
properties and point it to the downloaded file.
Addresses part of T7960
Test Plan: Manual, some specs
Reviewers: evan, spang, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4249
Summary: Using `await` instead of `advanceClock()` fixes all the things!
Test Plan: Ran the spec file
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4248
Summary:
This fixes T7995.
Previously, attempting to log out from your NylasID would just
automatically sign you back in with the same NylasID because the webview
session was preserved.
Now, instead of relaunching the windows, we restart the app to clear the
webview session
Test Plan: manual
Reviewers: halla, evan
Reviewed By: evan
Maniphest Tasks: T7995
Differential Revision: https://phab.nylas.com/D4224
Summary:
Previously, the logic for deciding wether to report an APIError lived
inside NylasAPIRequest and depended on an array of ignorable
statusCodes. However, when handling APIErrors, NylasAPIRequest would
convert special error codes (for offline errors) into a statusCode of 0,
which would automatically be ignored.
Given that the delta streaming connection didn't convert status codes to
0, it wouldn't actually ignore the error and end up reporting it,
flooding sentry.
This commit makes it so that that logic lives inside APIError, and
anyone who handles APIErrors can check wether to report them or not,
including the delta streaming connection
Test Plan: manual :(
Reviewers: spang, evan, halla
Reviewed By: evan
Differential Revision: https://phab.nylas.com/D4235
Summary:
Currently, when we auth an account for the first time in Nylas Mail (or
we blow away the database), the app is going to request transactions
since cursor `null` from the /delta/streaming endpoint and from the local-sync
delta observable, instead of requesting transactions since cursor `0`
This is due to a subtle bug with the use of default values when
destructuring an object. Our coded did the following:
```
const {cursor = 0} = this._state
```
Which at a glance seems correct. However, this will only work as
expected if `this._state` has the following shape:
```
{cursor: undefined}
```
And unfortunately, our `this._state` looked like this when authing an
account for the first time:
```
{cursor: null}
```
Which would make `cursor === null` instead of `0`.
This is because when using default values, null is considered an
intentional argument/value, as opposed to not passing any argument/value
(which will mean that the argument is undefined).
This was a regression introduced in d60a23c and 8bc2ec5
Test Plan: manual, will add regression test in upcoming diff
Reviewers: evan, spang, halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D4243
Summary:
Previously, after creating a new folder, the UI would indicate that the new
folder had children, even though it didn't. This was caused by duplicate models
in our `MutableQueryResultSet` for the user's categories. Basically, we would
sync the server version of the folder before the `SyncbackTask` for the new
folder returned its `serverId`. Without the `serverId`, the synced version of
the folder couldn't yet be tied to the optimistic folder, so a second row was
created in the database. This second row is removed when the `syncbackTask`
does return the `serverId`, because we persist the optimistic folder with a
`REPLACE INTO` query. (This deletes other rows with the same id.) However,
since this was done inside a `persist` change with the `serverId` and no
`unpersist` was ever recorded for the `clientId`, our `MutableQueryResultSet`
never removed the `clientId` model.
To address this, this diff adds a check in `updateModel` to see if the
`serverId` is being added. If it is, and both the `serverId` and `clientId`
exist in the `_ids` list, we remove the `clientId`.
The children indicator does still briefly show up while there are still two
separate rows for that folder in the database. If we want to get rid of this
completely, we would have to ensure that we do not sync the folder before the
`syncbackTask` returns the `serverId`. However, this would probably be pretty
involved, and for not much gain. This fix is much simpler and reduces most of
the issue.
Test Plan: manual
Reviewers: juan, evan
Reviewed By: evan
Differential Revision: https://phab.nylas.com/D4228
Summary:
The ignore list was very old. It included several dozen MB of docs_src and
other crap in our builds
Test Plan: manual
Reviewers: halla, spang, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D4213
Summary: We need to upload the nupkg for the Windows autoupdater to work
Test Plan: manual
Reviewers: juan, spang, halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D4219
Summary:
This is going to be a diff way shorter than the previous one! Basically, it adds a new endpoint, `/blobs` to our API to store send later attachments. When a user schedules a draft to be sent, we send all attachments to this endpoint. Separately, we store the rest of the message as metadata.
When it's time to send the message, we fetch the attachments from S3, fetch the metadata and merge them together to get a message we can send.
Test Plan: Tested manually. Will make a final QA pass before landing.
Reviewers: juan, halla, evan
Reviewed By: halla, evan
Differential Revision: https://phab.nylas.com/D4196
Summary:
On windows running `git rev-parse --short HEAD` does in fact now give you
9 characters instead of 7 like it does on Mac. This will ensure that
builds get uploaded to the same folder and help ensure we don't post a
version that doesn't exist on the release page.
Test Plan: Manual
Reviewers: juan, halla, spang
Reviewed By: spang
Differential Revision: https://phab.nylas.com/D4217
Summary:
Use electron's `powerMonitor` module to detect when the computer resumes
from sleep, and restart the sync loop when that happens in order to
sync the inbox immediately, in case we received any new mail events
while the computer was asleep
Test Plan: manual
Reviewers: evan, spang, halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D4216
Summary:
This commit makes it so `resetEmailCache` works as expected, i.e. it
removes all databases, without forcing the user to re sign-in to their
accounts or NylasID
Previously, this method removed the database without removing the
accounts, left users in an un-authed state that was hard to recover
from. This was fixed in D4212 which makes sure that when we get a new
identity, sync and deltas are restarted
However, resetEmailCache would still force you to log in to yoru NylasID
because it was deleted from the database. However, if we reuse the
command `application:relaunch-to-initial-windows` instead of manually
deleting the database, we can relaunc the app while preserving the users
NylasID session, so they don't have to sign back in manually.
Test Plan: manual
Reviewers: evan, spang, halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D4215
Summary:
This prevents the app from entering a restart loop when there's no
identity.
Specifically, when a user logs out of their identity, or when she resets the
email cache, or any other scenario that leaves the app without an
identity but with accounts added, the sync loop (and deltas) will start
without an Identity.
This will cause NylasAPIRequest to throw an error that
forces the user to close the app. When the app restarts, sync will start
again without an identity, and the user will be forced to close the app
again, and so on and so forth for the rest of eternity
Relevant error:
https://github.com/nylas/nylas-mail-all/blob/master/packages/client-app/src/flux/nylas-api-request.es6#L165-L174
Additionaly, this makes sure that after resetting the email cache, the sync
process starts when the identity becomes available
This solves T7989
Test Plan: manual
Reviewers: evan, spang, halla
Reviewed By: halla
Differential Revision: https://phab.nylas.com/D4212
Summary:
build-client, aka packaging, aka bundling is now separate from uploading.
This is both to compartmentalize our tasks a bit more and so we can add a
non-grunt windows task inbetween packaging and uploading.
No more heavily-overloaded PUBLISH_BUILD flag.
Added SIGN_BUILD flag instead.
No more TRAVIS and TRAVIS_PULL_REQUEST flag.
Test Plan: Manual
Reviewers: halla, juan, spang
Reviewed By: juan, spang
Differential Revision: https://phab.nylas.com/D4208
Summary:
This removes specs from the build production version.
One of the major issues with the windows build was that one of the package
specs included a .node compiled file that the codesigner attempte to sign,
but couldn't find for some reason.
In testing, removing the specs from the build prod version fixed this.
While we could just remove the one offending test, I think we should just
remove the whole suite from the build version. That'd save in the number
of files we have to ship, save download time, save build time, and have
less chance for some symlinked issue failling a build in the future.
Test Plan: Manual
Reviewers: khamidou, halla, juan, spang
Reviewed By: juan, spang
Differential Revision: https://phab.nylas.com/D4206
