$ npm install tiny-lr-fork
This is a fork of tiny-lr. The maintainer of tiny-lr appears to be on a hiatus and this fork takes care of lingering issues until the maintainer of tiny-lr (hopefully) returns.
Changes made:
console.log
sThis script manages a tiny LiveReload server implementation you can spawn in the background.
It exposes:
a background-friendly bin wrapper (thanks to @FGRibreau pid.js gist)
Grunt tasks to start the server and trigger reload notification. Every task
name is prefixed by tinylr-
.
Generic targets to include in
your Makefile (include node_modules/tiny-lr/tasks/tiny-lr.mk
)
It doesn't have any watch ability, it must be done at the build process or application level.
Instead, it exposes a very simple API to notify the server that some changes have been made, that is then broadcasted to every livereload client connected.
# notify a single change
curl http://localhost:35729/changed?files=style.css
# notify using a longer path
curl http://localhost:35729/changed?files=js/app.js
# notify multiple changes, comma or space delimited
curl http://localhost:35729/changed?files=index.html,style.css,docs/docco.css
Or you can bulk the information into a POST request, with body as a JSON array of files.
curl -X POST http://localhost:35729/changed -d '{ "files": ["style.css", "app.js"] }'
As for the livereload client, you need to install the browser extension: http://feedback.livereload.com/knowledgebase/articles/86242-how-do-i-install-and-use-the-browser-extensions- (note: you need to listen on port 35729 to be able to use with your brower extension)
or add the livereload script tag manually: http://feedback.livereload.com/knowledgebase/articles/86180-how-do-i-add-the-script-tag-manually- (and here you can choose whatever port you want)
This package exposes a bin
you can decide to install globally, but it's not recommended.
tiny-lr --help
Usage: tiny-lr [options]
Options:
-h, --help - Show help usage
-v, --version - Show package version
-p, --port - Port to listen on (default: 35729)
--pid - Path to the generated PID file (default: ./tiny-lr.pid)
The best way to integrate the runner in your workflow is to add it as a reload
step within your build tool. This build tool can then use the internal binary
linked by npm in node_modules/.bin/tiny-lr
to not rely on global installs (or
use the server programmtically).
You can start the server using the binary provided, or use your own start script.
var tinylr = require('tiny-lr');
// standard LiveReload port
var port = 35729;
// tinylr(opts) => new tinylr.Server(opts);
tinylr().listen(port, function() {
if(err) {
// deal with err
return;
}
console.log('... Listening on %s (pid: %s) ...', port);
})
You can define your own route and listen for specific request:
var server = tinylr();
server.on('GET /myplace', function(req, res) {
res.write('Mine');
res.end();
})
And stop the server manually:
server.close();
This will close any websocket connection established and emit a close event.
To use as a connect / express middleware, tiny-lr needs query / bodyParse middlewares prior in the stack.
Any handled requests ends at the tinylr level, not found and errors are nexted to the rest of the stack.
// This binds both express app and tinylr on the same port
var app = express();
app.use(express.query())
.use(express.bodyParser())
.use(tinylr.middleware({ app: app }))
.use(express.static(path.resolve('./')))
.use(express.directory(path.resolve('./')))
.listen(35729, function() {
console.log('Listening on %d', 35729);
})
The port you listen on is important, and tinylr should always listen on
the LiveReload standard one: 35729
. Otherwise, you won't be able to rely
on the browser extensions, though you can still use the manual snippet
approach.
You can also start two different servers, one on your app port, the
other listening on the LiveReload port. Check the
examples/express/server.js
file to see how.
This package exposes a tasks/
directory, that you can use within your Gruntfile with:
grunt.loadNpmTasks('tiny-lr');
tinylr-start
should be used with the watch
task, probably with an alias
that triggers both tinylr-start watch
tasks.
tinylr-reload
should be configured as a "watch" task in your Gruntfile.
grunt.initConfig({
watch: {
reload: {
files: ['**/*.html', '**/*.js', '**/*.css', '**/*.{png,jpg}'],
tasks: 'tinylr-reload'
}
}
});
grunt.registerTask('reload', ['tinylr-start', 'watch']);
See tasks/tiny-lr.mk
.
Include this file into your project Makefile to bring in the following targets:
Then define your "empty" targets, and the list of files you want to monitor.
CSS_DIR = app/styles
CSS_FILES = $(shell find $(CSS_DIR) -name '*.css')
# include the livereload targets
include node_modules/tiny-lr/tasks/*.mk
$(CSS_DIR): $(CSS_FILES)
@echo CSS files changed: $?
@touch $@
curl -X POST http://localhost:35729/changed -d '{ "files": "$?" }'
reload-css: livereload $(CSS_DIR)
.PHONY: reload-css
The pattern is always the same:
touch
the directory to update its mtimelivereload
and the list of files to "watch" as
prerequisitesYou can chain multiple "reload" targets in a single one:
reload: reload-js reload-css reload-img reload-EVERYTHING
Combine this with visionmedia/watch and you have a livereload environment.
watch make reload
# add a -q flag to the watch command to suppress most of the annoying output
watch -q reload
The -q
flag only outputs STDERR, you can in your Makefile redirect the
output of your commands to >&2
to see them in watch -q
mode.
npm test
var url = parse(this.request.url);
var server = this.app;
var ws = this.ws = new WebSocket('ws://' + url.host + '/livereload');
ws.onopen = function(event) {
var hello = {
command: 'hello',
protocols: ['http://livereload.com/protocols/official-7']
};
ws.send(JSON.stringify(hello));
};
ws.onmessage = function(event) {
assert.deepEqual(event.data, JSON.stringify({
command: 'hello',
protocols: ['http://livereload.com/protocols/official-7'],
serverName: 'tiny-lr'
}));
assert.ok(Object.keys(server.clients).length);
done();
};
properly cleans up established connection on exit.
var ws = this.ws;
ws.onclose = done.bind(null, null);
request(this.server)
.get('/kill')
.expect(200, function() {
console.log('server shutdown');
});
# tiny-lr
## GET /
respond with nothing, but respond.
request(this.server)
.get('/')
.expect('Content-Type', /json/)
.expect('{"tinylr":"Welcome","version":"0.0.1"}')
.expect(200, done);
unknown route respond with proper 404 and error message.
request(this.server)
.get('/whatev')
.expect('Content-Type', /json/)
.expect('{"error":"not_found","reason":"no such route"}')
.expect(404, done);
## GET /changed
with no clients, no files.
request(this.server)
.get('/changed')
.expect('Content-Type', /json/)
.expect(/"clients":\[\]/)
.expect(/"files":\[\]/)
.expect(200, done);
with no clients, some files.
request(this.server)
.get('/changed?files=gonna.css,test.css,it.css')
.expect('Content-Type', /json/)
.expect('{"clients":[],"files":["gonna.css","test.css","it.css"]}')
.expect(200, done);
## POST /changed
with no clients, no files.
request(this.server)
.post('/changed')
.expect('Content-Type', /json/)
.expect(/"clients":\[\]/)
.expect(/"files":\[\]/)
.expect(200, done);
with no clients, some files.
var data = { clients: [], files: ['cat.css', 'sed.css', 'ack.js'] };
request(this.server)
.post('/changed')
.send({ files: data.files })
.expect('Content-Type', /json/)
.expect(JSON.stringify(data))
.expect(200, done);
## GET /livereload.js
respond with livereload script.
request(this.server)
.get('/livereload.js')
.expect(/LiveReload/)
.expect(200, done);
## GET /kill
shutdown the server.
var server = this.server;
request(server)
.get('/kill')
.expect(200, function(err) {
if(err) return done(err);
assert.ok(!server._handle);
done();
});
© 2010 - cnpmjs.org x YWFE | Home | YWFE