For my team at work, I\'m trying to set up a semi-automated JavaScript script and dependency management system with the help of Gulp and Browserify.
Figured it out, sharing the learns:
Code example:
var gulp = require('gulp'),
source = require('vinyl-source-stream'),
browserify = require('browserify'),
factor = require('factor-bundle');
gulp.task('browserify', function(){
return browserify({
entries: ['blog.js', 'page.js']
})
.plugin(factor, {
// File output order must match entry order
o: ['bundle/blog.js', 'bundle/page.js']
})
.bundle({
debug: true
})
.pipe(source('common.js'))
.pipe(gulp.dest('bundle/'));
});
The key difference between this output and the diagram, is that the common.js
file is automatically generated based on common depenedencies between blog.js
and page.js
. This is described in the factor-bundle documentation.
Notes:
I found that Node would throw an error if the output files didn't already exist. I'm unsure why as I would have assumed that factor-bundle would simply write a stream to the outputting file regardless of whether it was there or not. As a workaround, after 'cleaning' the output directory, I simply created 'placeholder' files to keep it happy.
I haven't figured out how to access the factor-bundle stream event when using it as a browserify plugin (it may not even be possible), so any further work on the output files (such as uglifying etc) would likely need to be done somewhere else in the pipeline, possibly with another browserify plugin, or even after the fact.