On Mon, Nov 27, 2017 at 11:15 AM Eli Zaretskii wrote: > The "second file" is .dir-locals-2.el, i.e. the value. > Correct. But the current state of documentation would lead one to believe that the second file name is ".dir-locals-2.el" because of the dir-locals-file-2 constant; it's actually derived from dir-locals-file only. So maybe we need to add that to the manual. > I have tried to clarify this in the attached patch. But its (constant) value does override, doesn't it? > ".dir-locals-2.el" file does override ".dir-locals.el". But that would happen even if the value of dir-locals-file-2 were "foo.el" i.e. at present dir-locals-file-2 value is completely unused.. it's just a documentation holder. I have now moved that documentation to the doc string of dir-locals-file constant. Which change? You are arguing about multiple issues, and I'm confused > about what change do you have in mind, exactly. > Apologies for the confusion, looking at NEWS, I learned about dir-locals-file-2 constant, but then grepping for that variable led me nowhere. So I initially thought that that constant was a relic of now-removed feature. I was wrong about that part.. the feature is still there and just that dir-locals-file-2 was not used. So the "change" which I propose in the attached patch is remove dir-locals-file-2 constant and put the documentation about this feature in the right places. > That variable is anyways not used.. so what's the point of keeping > docstring of an unused > > variable. > > The doc string documents a feature. If you remove the doc string and > do nothing else, where will that feature be documented? > Yes, there was miscommunication.. I proposed to remove that doc string and add references to ".dir-locals-2.el" wherever applicable ( http://lists.gnu.org/r/emacs-devel/2017-11/msg00568.html). Then let's fix the documentation. > I didn't say it was a problem, you misunderstood what I said. Feel > free to submit such a patch, and thanks in advance. > Thanks! I have attached the patch. -- Kaushal Modi