Experiencing extreme Finder hang when rebuild happens about ember-app-kit

[stefanpenner]

This sounds mega painful. TBH I have totally disabled spotlight, so maybe there is a glaring issue I just have not noticed.

That being said, AFAIK we only lock a single file.

This seems like it needs some more investigation, are there any other details you can provide? Maybe an example project that demonstrates this issue?

Are these files also being observed by something else, like dropbox?

[rwjblue]

We are only locking one file, but if you are running grunt test:server or grunt server it takes a bit of CPU (2.5% on my retina macbook pro). I believe the reason for this is the watch task is polling the disk for changes (this is dependent on the platform you are running on of course).

This may be related (or not): gruntjs/grunt-contrib-watch#35

[stefanpenner]

seems crazy that it would be polling on OSX.

Also is this by any chance also observing some temp directories, something like .sass-cache or something, as that might also explain it.

[rwjblue]

seems crazy that it would be polling on OSX

Agreed. More likely it's tangentially related (some other process Dropbox, backup, etc) and this compounds the issue.

@jayphelps - SSD?

[jayphelps]

@stefanpenner @rjackson Looking into this more tonight.

SSD: yes.

fs.watchFile uses kqueue BSD system call for updates (last time I checked), no polling. (assuming EAK uses watchFile)

[jayphelps]

This was originally happening with the example app. Tried things again and of course...issue is gone. As suggested, most likely related to some other service trying to battle it out. Also--I'm using Total Finder which uses mach_star to inject code into the Finder process. Don't want to jump the gun with circumstantial evidence but it's something to consider.

Feel free to close, unless someone else can reproduce. I've added more information about my system to the top, for future reference in case others have issues and seek correlation. No troubles at all on my iMac either.

[stefanpenner]

@jayphelps thanks for the follow up. As suggested I'll close for now, but I will keep an eye on it.