During the Meteor 1.6.1 beta period, we introduced logic to render a
<script> tag to load the SockJS library in older browsers (#9353), and so
it seemed important to run test-packages both with and without the
<script> tag, using a special query parameter appended to the app URL.
The #9353 changes were ultimately reverted before Meteor 1.6.1 was
released (see 365804218f), and Meteor 1.6.2
will take a very different approach to bundling dependencies like SockJS
for legacy browsers (#9439). As part of this approach, PhantomJS is always
considered a legacy browser, and as such provides valuable feedback on the
behavior of web.browser.legacy bundles. However, since there's nothing to
configure with regard to SockJS anymore, there's no point in running the
test-packages suite twice in PhantomJS.
In order to run these tests in a modern browser environment, we should
probably revisit the idea of running tests in headless Chrome:
https://github.com/meteor/meteor-feature-requests/issues/254
This could have prevented #9631 from being a problem in the first place,
and feels like the right thing to do, considering the number of cases
we've seen of people accidentally including babel-preset-meteor in their
.babelrc files.
As usual, we have to bump the ecmascript version as well, so that the
compiler plugin will be rebuilt.
In the conversion to ECMAScript classes in
https://github.com/meteor/meteor/commit/6fcf190bb489db19, it appears that
JSDoc `@method`, `@instance` and `@memberof` declarations were already
in place for the (outgoing) `Mongo.Collection.prototype.{foo}` expressions,
with the exception of `upsert`. The omission caused docs generation failure
with the new ES syntax since, presumably, it was less possible to be inferred
from the `MCp*` notation.
Additionally, this changes the casing of `@memberOf` to `@memberof`,
mainly because:
* The JSDocs use this notation.
* The data JSON outputted by JSDoc lowercases property keys, including
cases where `memberOf` was capitalized, making it less clear that
there is a 1:1 mapping.
* My editor refuses to syntax-highlight jsdoc declarations with [A-Z],
possilby because of the above reasons.
Changes:
* Use ECMAScript class and arrow function syntax for Tracker.Computation
and Tracker.Dependency.
* Remove the deprecated.js file (no more Deps, Tracker.depend, or
Tracker-related Meteor.* APIs).
Should fix#9598, thanks to @abernix's diagnosis of the problem:
https://github.com/meteor/meteor/issues/9598#issuecomment-361570289
My later comment in that issue thread is not accurate, since the Babel
wrapNativeSuper helper already pulls in the core-js Reflect.construct
polyfill. Instead, the root of the problem really seems to be Babel's
generation of an unguarded _typeof(Reflect) expression.
This reverts commit dfc0702558.
We've seen some odd test failures (e.g. `passwords - tokens`) and trouble
updating from 1.6.1 to 1.6.2-beta.3 (easily solved with `meteor reset`,
but also worth investigating), so I think we should keep working on the
Mongo 3.6 upgrade in a PR, rather than running all of our tests against
a devel branch that includes Mongo 3.6.
cc @abernix @hwillson
The change made in a30f42c4ac switched
from CommonJS to ECMAScript export notation.
It seems JSDoc isn't smart enough to make the same association as it did with
the previous notation as it does for an anon. function expression `export`-ed
as a `const`ant. We could annotate this with `@function check`, but it seems
reasonable to just export a function declaration directly, which JSDoc will
understand.
This, along with ababb18aef, corrects the
failure of docs generation originally discovered in:
8f7ceb5de5.
Refs: https://github.com/meteor/meteor/pull/9593
It's nice to match versions, but we couldn't bump the minor version of the
Meteor jquery package without breaking compatibility with the current Meteor
release (1.6.1), since jquery is a core package.
The next Meteor release (either 1.6.1.1 or 1.6.2) will either use the new
1.12.1 version or (ideally) jquery will no longer be a core package, and
thus will not be constrained by the release.
Related: #9605
Note that you can run `meteor npm install jquery` (any version) in your
top-level application directory, and the meteor/jquery package will use
that version instead of 1.12.1.
The changes made in 8bbfd531c1 prevent
JSDoc (used to generate docs for https://docs.meteor.com via the
https://github.com/meteor/docs repository) from associating the
convenience methods which are exposed on the `Meteor` namespace
(e.g. `Meteor.call`, `Meteor.apply`, `Meteor.reconnect,` etc.) as they are
being identified as members of the `Connection` class. While this is
techincally true, and all of these are also available on
`Meteor.connection` (e.g. `Meteor.connection.status`), we have
historically exposed them in the docs using their (preferred) `Meteor.*`
aliases.
Hopefully, this at least partially resolves
8f7ceb5de5.
This reverts commit b06a6af335, where I
attempted to allow `Mongo.Collection` to be subclassed by non-native class
syntax by implementing static `call` and `apply` methods.
As the reproduction in #9595 demonstrates, in order to subclass
`Mongo.Collection` properly given this trick, one would need to override
not only the `constructor` method but also the `init` method that the
`Mongo.Collection` base class calls. Since this implicit expectation is
too much to ask of new subclassing code, let alone existing subclassing
code, I'm afraid the only remaining option is to go back to implementing
`Mongo.Collection` as a traditional constructor function, rather than
using native `class` syntax, so that subclasses can invoke the constructor
using `Function.prototype.{call,apply}` as they've always done.
Changes:
* Use ECMAScript class and arrow function syntax for Tracker.Computation
and Tracker.Dependency.
* Remove the deprecated.js file (no more Deps, Tracker.depend, or
Tracker-related Meteor.* APIs).
Because Promise.asyncApply is only defined on the server, and Meteor 1.6
no longer uses Regenerator to compile async/await and generator functions
on the server, this code no longer serves any meaningful purpose.
On top of that, the babel-runtime.js module is loaded on the client, so
this code was forcing the Regenerator runtime to be included in the client
JS bundle, even if generator functions were not used anywhere else in the
application.
The benefit of removing this @babel/runtime/regenerator dependency won't
be fully apparent until Meteor 1.6.2, since there are probably other
places in client code that depend on it, so it will probably still be
bundled in most applications. However, the new web.browser.legacy system
(#9439) should remove most of those dependencies for modern browsers, as
Meteor 1.6.2 will no longer use Regenerator to compile async/await and
generators for the modern JS bundle.
Update the Meteor Tool to use Mongo 3.6.2 for 64-bit OS'
and Mongo 3.2.18 for 32-bit OS'. A few important mentions:
- As of Mongo 3.6, all Mongo binary downloads include
SSL - there is no longer a non-SSL based download bundle
(so it's a bit bigger, but that shouldn't be an issue).
- Using the `--nojournal` option with WiredTiger based
replica sets is no longer supported (see
https://jira.mongodb.org/browse/SERVER-30760). The
`--nojournal` flag was added in
bcfe072d52
to help reduce the amount of disk space used by Mongo,
but since this option is no longer supported, we'll
have to live with the extra disk space usage.
- Add PR link to History.md
Thanks to bug reports from @jamesmillerburgess, in Meteor 1.6.1-rc.4 we
reverted a change that unsuccessfully attempted to remove SockJS from the
JS bundle (and instead injected a <script> tag to load SockJS in older
browsers): 365804218f
With this commit, we are once again attempting to exclude SockJS from the
JS bundle for modern browsers, this time taking advantage of the
web.browser.legacy system to control which browsers receive which bundle.
The definition of a "modern" browser is stricter now (essentially, any
browser with native async/await support), so I'm hopeful that browsers
with incomplete WebSocket support will get the legacy bundle, which
includes SockJS.
However, if there are still bugs in modern browsers due to faulty
WebSocket message compression (for example), we should be able to address
those problems directly, rather than reverting this change yet again.
I missed these packages when publishing 1.6.2-beta.0, since I hadn't
bumped their versions in #9439 as I should have, so the release script did
not republish them. Because of this, 1.6.2-beta.0 will not be usable, and
so I will publish 1.6.2-beta.1 ASAP.
The boilerplate-generator version bump is patch rather than minor, since
the relevant changes to that package seemed backwards-compatible.
I've decided to stick with Meteor.isModern for now, since it's false for
both web.browser.legacy and web.cordova bundles, which would make logic
involving a hypothetical Meteor.platform property more complicated than
simply using Meteor.isModern. I'm open to revisiting this later.
