Although I think 8c70716954 is a good idea
in practice, it altered the timing of self-tests enough to cause a number
of failures, so for now that behavior will be gated behind the environment
variable METEOR_WATCH_PRIORITIZE_CHANGED.
This drastically reduces the number of open file descriptors by not
preemptively acquiring a file descriptor for every watched file.
The downside is that the first time you edit a particular file, you may
have to wait up to DEFAULT_POLLING_INTERVAL milliseconds before Meteor
notices the change. The upside is that changes will be detected
instantaneously for that file in the future, even after restart.
If holding open too many file descriptors is indeed a contributing factor
to issues such as #8648, then this change should go a long way towards
mitigating those problems.
cc @abernix @hwillson
Render callbacks can now inject HTML content into multiple different
elements, and may also append content to the <head> or <body> elements, on
both the client and the server.
This new API was inspired by trying to use the styled-components npm
package on the server, which involves not only rendering and injecting
static HTML somewhere in the <body>, but also appending the resulting
<style> tag(s) into the <head>:
import { onPageLoad } from "meteor/server-render";
import { renderToString } from "react-dom/server";
import { ServerStyleSheet } from "styled-components";
onPageLoad(sink => {
const sheet = new ServerStyleSheet();
const html = renderToString(sheet.collectStyles(
<App location={sink.request.url} />
));
sink.renderIntoElementById("app", html);
sink.appendToHead(sheet.getStyleTags());
});
Note that the server-render package now exports an onPageLoad function,
rather than the old renderIntoElementById function. The functionality of
renderIntoElementById is now exposed by the {Client,Server}Sink API.
I say the client-side version of this API is 'isomorphish' to the
server-side version, because ClientSink methods can accept DOM nodes in
addition to raw HTML strings, whereas DOM nodes don't really make sense on
the server.
The server-render package requires webapp@1.3.17 or later, but using a
non-prerelease version contraint for a package involved in the release
(i.e., webapp) is tricky during the prerelease phase, since the -beta.n
version is strictly enforced.
This API allows registering callbacks that have the opportunity to modify
boilerplate.baseData on each request, which will be useful for
implementing server-side rendering.
The code in question behaves the same as before if there are no callbacks
registered, so this change should be completely backwards compatible.
As discussed in https://github.com/meteor/meteor/pull/8777 it seems best
to always have the suffix on npm-wrapper packages.
If anything, as a reminder when bumping the version on wrapper packages,
but also just to generally make it more clear if a version suffixed with
`_1` is actually higher than a non-suffixed version or not.
