Closed odoe closed 8 years ago
I'd vote for not using browser-sync, or at least be sure to test it on Windows (I can help w/ that). I've been having a lot of trouble in other projects w/ browser-sync on Windows - sometimes taking up to 30 seconds to serve a page (after any transpilation, pre-processing etc has happened). Actually, I'm pretty sure I had these problems w/ http-server in test apps I built from this generator too at one point (didn't we talk about that before?), but I can't reproduce them now.
All I can say for sure is that browser-sync is unusable for me in other projects, and that I've never had any problems w/ grunt-contrib-connect (which also supports livereload w/o manual script insertion, even on projects where we're loading local esri/dojo packages.
If you're close on BrowserSync and want to throw up a branch/PR for me to test on Windows I can. Otherwise, would you be interested in a PR using connect-livereload instead?
Yeah, at this point, I think I'll leave the injected livereload script as-is. BrowserSync has been a little slower in my testing, so no changes for now.
I've been testing grunt-browser-sync as a replacement for
http-server
andlivereload
. It would save the step of having to inject thelivereload
script into the page during dev and so far has been working petty well.Will put it through a few more paces before making the change,