Summary:
Adds a fun new UI for adding accounts to the sync engine. After creating your sync engine instance, all you have to do is auth your accounts on the command line and then enter the URL/port number in this flow. That pulls all of your accounts from the `/accounts` endpoint, mocks an identity token, and edits your `config.json` properly.
TODO: Update the docs in the repo and revert the PR with the temporary fix.
Test Plan: Tested locally.
Reviewers: bengotow, halla, juan
Reviewed By: halla, juan
Differential Revision: https://phab.nylas.com/D3114
Summary:
Prevent hitting SMTP rate limit with multi-send
Some users were hitting the SMTP rate limit when sending messages
with multi-send to too many recipients. There is a backend patch
to help with this, but we decided it would be useful to limit it
on the client side as well.
Test Plan: Added unit tests
Reviewers: jackie, juan
Reviewed By: jackie, juan
Differential Revision: https://phab.nylas.com/D3109
Summary:
Extended default signatures to also be associated with aliases in the signature
preference page. Also fixed some styling with the signature dropdown button in its blurred
state on the popout composer.
fix(signatures): Fixed styling for signature compuser button on popout when blurred
test(signatures): Updated tests to support and cover extending signatures to alias
Test Plan: Modified existing tests to work with these changes
Reviewers: juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3111
Summary: We used to parse the quoted text on each keystroke in the composer for a reply so that we could continue to determine what was quoted text. However, that resulted in dramatically slow typing for replies to complex HTML emails. Now, the quoted text isn't a part of the reply until `prepareDraftForSyncback`, after all of the extensions have run their transformations—we use a marker to determine whether quoted text should be appended or not. The quoted text control is now a one-way operation—you can't hide the quoted text after showing it (Gmail-style).
Test Plan: Tested locally (but didn't run unit tests because they won't run on my machine...)
Reviewers: bengotow, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3106
Summary:
refactor(test): Fix file paths in message item body tests to be prepended with file://
Some inline images were not rendering as seen on (#621). The images were being correctly added to the
downloads folder, however, their path was being prepended with an unrelated base url. I hardcoded
file:// into the image paths so that they would always point to the local downloads folder.
Test Plan: Ran the specific tests for message item body and message item. Also checked in my email.
Reviewers: juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3105
Summary: Some emails were having their body height set to 0 if their height was dependent upon the iframe height -- e.g. body.height equals 100% or inherit. Added a check to see if height computed to 0px and if so, set the height to auto.
Test Plan: Tested with emails on my computer
Reviewers: juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3100
This modifies the behavious when pasting into the to, cc or bcc fields of a new draft.
Any text already in the input field is not automatically deleted. Rather the pasted
text is added to the already existing text. If the new contents of the email field now
matches as an email address (based on RegExpUtils.emailRegex().test()), then the address
is immediately tokenized. If the new input text doesn't match as an email, just append
the new text and wait for further input.
When APIRoot is using https, the port will forced to be 443 even when APIRoot
indicates another port. This is because of the line options.port = 443
overwriting the correct options.port extracted by url.parse. This is fixed by
removing the problematic line. The default port is still 443 if you look into
https.js. In case we want a different default port in the future, we can do
options.port = options.port || 443.
Summary:
Refactored signature preferences page to allow more signatures than the previous 1-1 mapping for signatures and accounts. Created a multi select dropdown of the accounts for which a certain signature is set as default for. Added a button into the draft header From field to toggle between saved signatures.
refactor(signatures): Add basic add/remove capabilities to static
refactor(signatures): Hooked up signature actions and signature store for basic functionality
fix(signatures): Cleaned up signature store and started on multiselectdropdown
fix(signatures): Add multi signature toggle select to multiselect dropdown
build(signatures): Built framework for multiselect dropdown
build(signatures): Toggle button functionality for dropdown
build(signatures): Build multi select from components and add debounce
refactor(signatures): Move signature actions and signature store into flux
fix(signatures): Styled composer signatures button/dropdown and fixed preferences checkmarks
build(signatures): Finish main functionality, about to refactor composer signature button into injected component
fix(signatures): Changed position styles
fix(signatures): Fixed background color for dropdown button when blurred
build(signatures): Began to write tests for signatures store, preferences and dropdown
Test Plan: Wrote tests for preferences signatures, signature store, signature composer dropdown and refactored tests for signature composer extension. For signature composer extension I removed applyTransformsToDraft and unapplyTransformsToDraft and tested by sending emails with signatures to different providers to make sure the <signature> tag caused problems.
Reviewers: bengotow, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3073
Fixes#1393
The objects returned by `new Notification()` have a `.close()` method.
So we can just keep a list of notifications that have been sent, and
close them when that message has been read.
Conflicts:
internal_packages/unread-notifications/lib/main.es6
* Remove double negative from test label
It always takes me a second to read a double negative, and be sure I
understood correctly. It's also particularly troublesome for non-native
English speakers.
* Change pending unread notification stack to be a queue
Consider the case where a user launches N1, and has 3 unreads since
their last launch. Additionally, after N1 has launched, they begin
receiving a new mail every 2 seconds.
Current behavior would be a notification for 1 of the 3 unread emails,
then every 2 seconds a notification for the most recently received mail.
They would only finally receive notifications for the other 2 original
messages once they stopped getting new mail.
* Fix private key email-adder, add "no private key" error
The decrypt UI is seriously confusing some people. This commit
adds an error message that should at least stop them from trying
to decrypt a message without a private key to speak of. Also,
there was a dumb hardcoded true in validAddress.
* Adds incorrect passphrase notification; enables carriage return for popover
The passphrase popover was woefully inadequate. It didn't tell users
when they had the wrong password - it just closed without saying anything -
and you couldn't even use carriage return to submit the password.
This commit fixes those mistakes by buffing out passphrase-popover.cjsx.
* Adds private key popover to decrypt button
The decrypt UI was confusing and didn't provide the user with
an option to get a key imported from the message view. This
mondo commit adds an entirely new popover so that the user
never again will be forced to go to the preferences page.
* Adds more forgiving encrypted block parsing
* Overhauls decryption error handling
The decrypt UI didn't clearly communicate error messages from the
failure in PGPKeyStore.decrypt up to the user. This commit adds
nice error surfacing as well as some pretty colors.
* Fix encrypt modal key miscount error via getKeyContents coercion
On Linux and Windows, fs.watch double-triggers on some actions in
the key folder, for reasons that aren't super clear. This was causing
issues where the encrypt modal would report not having a key loaded
even though the key was totally loaded and saved. This commit sort
of kludgily forces the modal to run getKeyContents for every key
it has saved right before it returns them all. This would probably
be better fixed with a refactor the the PGP Keystore.
* remember to close popover, d'oh
* patch key picker modal styling for Linux and Windows
* response to review
When running against a custom Sync Engine, viewing the onboarding page
on startup does not make much sense. This here adds code that checks for
the `env` of the Nylas config and skips the onboarding if it is `local`
or `custom`.
The clean script fails since productName is "Nylas N1" which contains a
space and it appears to pkill 2 patterns. This is causing the script to
stop in my linux machine. Now the productName is wrapped with quotes to
prevent this from happening.
Summary:
when renaming nested labels, maintain nested path
fixes#2402
Test Plan: created nested labels and successfully tested in dev
Reviewers: juan
Reviewed By: juan
Subscribers: evan
Differential Revision: https://phab.nylas.com/D3028
Summary: Specs were failing on Linux, this diff fixes them.
Test Plan: Unit tests
Reviewers: evan, juan
Reviewed By: juan
Differential Revision: https://phab.nylas.com/D3029
