Occasionally lineman stops noticing file changes.

[joegaudet]

It's non deterministic tho, so... I don't rreally have much else to say - happens more frequently after one of the pre processors eats shit cause of a syntax error.

[davemo]

I wonder if this is related to grunt-watch-nospawn; going to test and see how feasible it is to pull in grunt-watch-contrib now that gruntjs/grunt-contrib-watch#25 is fixed. (I have no idea if this will fix the issue, as you said it's non-deterministic, but it's worth a shot) :P

[searls]

Dave, note that grunt-contrib-watch has not been released since fixing that issue.

On Jul 12, 2013, at 4:51 PM, David Mosher notifications@github.com wrote:

I wonder if this is related to grunt-watch-nospawn; going to test and see how feasible it is to pull in grunt-watch-contrib now that gruntjs/grunt-contrib-watch#25 is fixed. (I have no idea if this will fix the issue, as you said it's non-deterministic, but it's worth a shot) :P

�
Reply to this email directly or view it on GitHub.

[davemo]

Oh, crap. Well that's no good :P

[jasonkarns]

grunt-contrib-watch 0.5.0, which includes the fix for gruntjs/grunt-contrib-watch#25, was released on 7/18. Should we consider switching from grunt-watch-nospawn to grunt-contrib-watch now?

In general, I very much like the idea of using the core contrib tasks when possible. Not to mention that we should be able to simplify the current watch config. (by utilizing the cascade behavior that supposedly works now in grunt-contrib-watch)

[jasonkarns]

Created grunt-contrib-watch branch and switched from grunt-watch-nospawn to grunt-contrib-watch. With options: spawn: false, all of our (limited) test suite passes.

EDIT: looks like it didn't pass TravisCI though 😦 https://travis-ci.org/testdouble/lineman/builds/10099566