robe-mode return lots of incorrect methods.

[dgutov]

robe-mode return lots of incorrect methods.

[dgutov]

When I using it for completing it always all methods for all objects.

That's fairly normal, unless the call target is statically known (and we don't do any complex analysis), completion will return methods from all classes. "Statically known" situations include calling a method on a class, or calling a method with implicit target.

Seeing that a call target is a constant (and using its class) should be relatively easy to implement, but that hasn't been done yet either.

However if I complete via C-M-i it will show "[No matches]".
Did I do something wrong?

It should show the same list of completions. What's the value of completion-at-point-functions in that Ruby buffer?

[madper]

Hi,
after I trigger completion-at-point it shows:
Contacting host: 127.0.0.1:24969 then shows "[No matches]".
BTW, shouldn't it only return String's methods when I call it after "a"..
BTW2. if I press C-M-i or call completion-at-point after "a".re, it will return 100 methods. But shouldn't it only return String's methods?

[dgutov]

You haven't answered my question. What's the value of completion-at-point-functions?

But shouldn't it only return String's methods?

Ideally, yes. Like I said, not implemented yet.

[madper]

Oops, I see. Thank you!
I'm using C-h v completion-at-point-functions and get following outputs:

completion-at-point-functions is a variable defined in `minibuffer.el'.
Its value is (robe-complete-at-point t)
Local in buffer test.rb; global value is 
(tags-completion-at-point-function)

[dgutov]

Hmm, okay. That looks normal.

When that happens, what's at the end of /tmp/robe-access.log?

[madper]

Hi, here is the log:

/tmp % cat robe-access.log 
I, [2014-10-12T00:05:41.230443 #18259]  INFO -- : GET /ping/
I, [2014-10-12T00:05:42.709240 #18259]  INFO -- : GET /complete_method/-/!/-/-
I, [2014-10-12T00:06:14.507644 #18259]  INFO -- : GET /complete_method/-/!/-/-

[dgutov]

Ok, again looks normal.

What's your installed versions of Emacs and Robe?

[madper]

(emacs-version) -> GNU Emacs 25.0.50.1 (x86_64-pc-linux-gnu, GTK+ Version 3.10.8) of 2014-10-10 on lgw01-11 ;;; does it too new?
robe -> 0.7.7 I installed it from marmalade via M-x list-package then choose robe to be installed.

[dgutov]

Ah. The Marmalade version should be pretty old now. Can you try it from MELPA?

[dgutov]

does it too new?

It's a bit too new. It's not exactly a problem, but you'd be better advised to use the 24.4 pretest until it comes out. There's not much shiny in Emacs 25 yet.

[madper]

The robe in MELPA is 20140915.1700. And I can reproduce the same issue with it. It's a bit odd that company-robe will always get lots of result. But C-M-i won't always work.

[madper]

Ah, maybe I should downgrade my emacs... :-/ thanks for your help. :-D

[dgutov]

Please let me know if it helps. Because I'm pretty sure the Emacs version shouldn't be the problem in this case.

[dgutov]

What about that unexpected token $end bit? Which package is it from?

[madper]

unexpected token $end should come with flycheck.
I tried to downgrade my emacs several days ago. But I got lots of error looks like:
no "function-put" I can't remember precisely...
I'll continue trying to downgrade it...

[rranelli]

I had problems downgrading my Emacs recently and recompiling the packages (i.e., removing the ~/.emacs.d/elpa directory altogether) solved the issue.

Maybe the errors you're facing are related to this.