Thanks! I was hoping it was something like that. I will add that now and test. I did look through the google closure docs for the annotations with the various outputs, but mostly followed whatever leads were already in the code and I did not look at the class level stuff I guess.
(I actually had a lot of head-scratching on release output at one point when I was working on the earlier stuff with the static bindable var output - until I found Josh's comment elsewhere in the code about using the 'deprecated' @expose instead of @export for static accessors) On Sat, Sep 3, 2016 at 2:52 AM, Alex Harui <aha...@adobe.com> wrote: > > > On 9/1/16, 11:40 PM, "Greg Dove" <greg.d...@gmail.com> wrote: > > >Took me a little while to get to this... > > > >Actually, I am really confused here. I don't have a real 'fix', but I > >understand what is causing it after chasing a few wild geese for a while. > >It seems the remove-circulars setting is the primary cause of this problem > >- removing that setting restores all the goog.requires stuff (which I > >*thought* I had already supported in the output), and after this js > >release > >mode seems fine. Is there any known bug here with remove-circulars? Or do > >I > >need to do something extra in the output to make it compatible with > >remove-circulars ? > > I forgot that it may not be as simple as adding goog.require. I just took > another look and saw that the constructor jsdoc is missing the @extends > (and/or @implements) directive which I think Google Closure Compiler uses, > but also the remove-circulars uses it to make sure that goog.requires for > base classes do not get removed. > > Hope that gets things working, > -Alex > >
