JSHint is about checking much more than whitespace and JSDoc (which is it doesn't check at all ;-) That is why we need both tools, IMHO.
I could write a whole lot about the tool and why it will certainly save your behind in large projects (check out my latest commits for examples), but the tool has an awesome webstite, that explains the purpose and features pretty good: http://www.jshint.com/about/ Check out the homepage as well, it's one of the most insane things I've seen on the interwebs in a long time: you can copy paste your JS code and it'll lint it on the fly and give you nice syntax highlighting so you can see what's going on. EdB On Thu, Nov 7, 2013 at 5:11 PM, Alex Harui <aha...@adobe.com> wrote: > Awesome, thanks. > > Any chance you can find time to put this 'recipe' in the wiki? > > What are some of the coding style differences you saw between files? Just > whitespace? Or more? Peter and I did run lint most of the time, but I > wasn't using --strict so maybe that's why things went off the rails. > > Thanks, > -Alex > > On 11/7/13 6:05 AM, "Erik de Bruin" <e...@ixsoftware.nl> wrote: > >>Ok, I have pushed the 'JSHint - gjslint' commits. >> >>The framework classes are now clean against this gjslint command: >> >>gjslint --strict --disable 0100 -r ./ >> >>I've used the --disable command because the new interface handling >>relies on a variable on the prototype that is initialised with a 'non >>primitive' value. >> >>I also ran the framework against a very loose set of JSHint tests >>using this 'header': >> >>/*jshint >> curly: false, >> eqeqeq: false, >> eqnull: true, >> globalstrict: true, >> indent: 2, >> maxlen: 120, >> onevar: false, >> strict: true, >> white: false */ >>/*global goog, org */ >> >>'use strict'; >> >>I had to use the loose set because the coding styles differ too much >>between files to allow stricter checks. I would suggest that we comply >>with both the above mentioned 'gjslint' arguments and with the >>following JSHint 'header': >> >>/*jshint >> globalstrict: true, >> indent: 2, >> maxlen: 120, >> strict: true, >> white: false */ >>/*global goog, org */ >> >>'use strict'; >> >>I suggest we use maxlen 120 to avoid a warning for long lines caused >>by long 'goog.require' statements. 'gjslint' takes care of the 80 char >>enforcement. >> >>EdB >> >> >> >>On Thu, Nov 7, 2013 at 8:57 AM, Alex Harui <aha...@adobe.com> wrote: >>> >>> >>> On 11/6/13 11:41 PM, "Erik de Bruin" <e...@ixsoftware.nl> wrote: >>> >>>>Alex, >>>> >>>>Do you also see the following warning when compiling DataBindingTest to >>>>JS: >>>> >>>>Data binding will not be able to detect assignments to 'strings'. >>>> >>>>dataProvider="{MyModel(applicationModel).strings}" /> >>> Yes. >>>> >>>>Just checking to make sure I have the same results as you... Also, >>>>when I run the app (debug and release), I see the string 'undefined' >>>>displayed several times on screen. Is that what you see as well? >>> No. Where do you see that? Did you get my changes? Databinding was >>> broken because the is/as code didn't handle "x as String". I put a new >>> FlexJSOverlay.zip on my people.a.o folder. Maybe you need that. >>> >>> -Alex >>> >> >> >> >>-- >>Ix Multimedia Software >> >>Jan Luykenstraat 27 >>3521 VB Utrecht >> >>T. 06-51952295 >>I. www.ixsoftware.nl > -- Ix Multimedia Software Jan Luykenstraat 27 3521 VB Utrecht T. 06-51952295 I. www.ixsoftware.nl
