Re[2]: [cedet-semantic] Completion with Xemacs not working

[Eric M. L. <er...@si...>]

>>> Nicolas Pernetty <nic...@ya...> seems to think that:
>> >> >2) if I type 'semantic-ia-complete-symbol-menu', a popup menu
>> >> >appears and when I click on the symbols, result is correct
>> >> >('var.tata' or 'var.titi') BUT if the completion is unique (for
>> >> >instance if I type 'var.ti') , I got an error : 
>> >> 
>> >> I fixed this with CVS version 1.9 of semantic-ia.el.
>> >
>> >Ok I've just tried and it doesn't seems to work (malformed list...),
>> >here is the stacktrace :
>> >
>> ># bind (standard-output stack-trace-on-signal debug-on-signal
>> >stack-trace-on-error debug-on-error)
>> >  delq(nil [("titi" variable (:type "int") (reparse-symbol
>> > classsubparts) [36 45])]) # bind (list title)
>> >  ad-Orig-imenu--create-menu-1("Completions" [("titi" variable
>> > (:type "int") (reparse-symbol classsubparts) [36 45])]) (setq
>> > ad-return-value (ad-Orig-imenu--create-menu-1 title list)) (if (and
>> > (null list) (not ...)) (setq ad-return-value "-") (setq
>> > ad-return-value (ad-Orig-imenu--create-menu-1 title list)))
>>   [ ... ]
>> 
>> Is this XEmacs?  I couldn't get a failure in Emacs 22.0.whatever
>
>Well I've just tried again with Xemacs, and this time activate 'debug
>on error', same error but with this you may have more information :
>
>Debugger entered--Lisp error: (malformed-list [("titi" variable (:type "int") (reparse-symbol classsubparts) [36 45])])
>  delq(nil [("titi" variable (:type "int") (reparse-symbol classsubparts) [36 45])])
>  ad-Orig-imenu--create-menu-1("Completions" [("titi" variable (:type "int") (reparse-symbol classsubparts) [36 45])])
>  (setq ad-return-value (ad-Orig-imenu--create-menu-1 title list))
>  (if (and (null list) (not ...)) (setq ad-return-value "-") (setq ad-return-value (ad-Orig-imenu--create-menu-1 title list)))
>  (let (ad-return-value) (if (and ... ...) (setq ad-return-value "-") (setq ad-return-value ...)) ad-return-value)
>  imenu--create-menu-1("Completions" [("titi" variable (:type "int") (reparse-symbol classsubparts) [36 45])])
>  imenu--mouse-menu((("titi : int" . [...])) #<buttondown-event button3> "Completions")
>  (let ((ans ...)) (when ans (if ... ...) (delete-region ... ...) (semantic-ia-insert-tag ans)))
>  (let* ((a ...) (syms ...) (pre ...)) (if (not syms) (error "No completions available")) (let (...) (when ans ... ... ...)))
>  semantic-ia-complete-symbol-menu(148)
>  call-interactively(semantic-ia-complete-symbol-menu)
>  command-execute(semantic-ia-complete-symbol-menu t)
>  execute-extended-command(nil)
>  call-interactively(execute-extended-command)
>
>Maybe I can do something to help you ?
>Add a print in the middle of the code for instance ? (I'm a real
>beginner in Elisp, so I couldn't figure it by myself...)

I looks like things are getting tied up in semantic-imenu.el around
line 55.  Perhaps David (who wrote that bit) has a suggestion?

[ ... ]
>> >> >a) I've read that there is some kind of auto checking while
>> >> >typing, resulting in underlining (color ?) what the parser doesn't
>> >> >understand. I've tried semantic-complete-analyze-inline ('inline
>> >> >completion not needed') and semantic-show-unmatched-syntax-mode
>> >> >(everything I type is in inverse video for 2 seconds, then go back
>> >> >to normal, even incorrect lines, for instance with mismatch
>> >> >parenthesis). Have I missed something ?
>> >> 
>> >> The mode is `semantic-show-unmatched-syntax-mode', but you
>> >> described `semantic-highlight-edits-mode'.
>> >> 
>> >> Unmatched syntax mode for C will not detect errors inside the body
>> >> of a function because function bodies are not (yet) parsed.
>> >
>> >Can you described an example where we should have something
>> >underlined ? Whatever I type, it doesn't change at all...
>> 
>> If semantic-show-unmatched-syntax-mode is on, then something like
>> this:
>> 
>> int int int main()
>> {
>> }
>> 
>> would underline some 'int's, but:
>> 
>> int main()
>> {
>>    int int int i;
>> 
>> }
>> 
>> would not.
>
>It doesn't work... Nothing happened at all, even if it's written
>'show-unmatched-syntax minor mode is enabled'

Use M-x list-faces-display (or the XEmacs equivalent) and make sure
that the face `semantic-unmatched-syntax-face' is visibly different
from the default.

  [ ... ]

Good Luck
Eric

-- 
          Eric Ludlam:                 za...@gn..., er...@si...
   Home: http://www.ludlam.net            Siege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: www.gnu.org