I tested directly to comment the lines you proposed and it fixes the issue.
If you want me to make some tests to reproduce/track the issue, feel
free to send me the procedure.
Eric M. Ludlam wrote:
> I currently suspect this may be related to a patch I made to
> semantic-symref-result-get-tags a little while ago. This function is
> now cleaning up files it opens after normalizing the tags that it
> Do you find that a buffer you have open gets deleted? If that were
> the case, then having other files pop open would make sense. What I
> don't understand is how the code in this function could let an already
> open buffer get killed.
> If you use
> M-x debug-on-entry RET semantic-symref-result-get-tags RET
> and then do your repro steps, then the first hit will show entering
> the function. Hit "c" to continue. If on the second hit the buffers
> get rearranged, then this would be the problem function.
> M-x cancel-debug-on-entry RET RET
> to clear that up.
> A quick fix for you would to be to go to line 399 and comment out the
> little block that delets buffers.
>>>> Damien Deville <damien.deville@...> seems to think that:
>> Hi Eric,
>> i tried your fix with only idle-summary-mode on and it does not seem to
>> fix the issue, when the bug arise running semantic-debug-idle-function
>> function does not provide any information at all in *scratch* nor
>> *Messages* buffer
>> Here is the way i intialize cedet in case it could be linked to my
>> config. I also use various packages that rely on cedet i tried without
>> them (company-mode, ca2, ...) and the issue is still there.
>> ;;--------------------- Load CEDET
>> (load-file "~/.emacs.d/lib/cedet-cvs/common/cedet.elc")
>> ;; use EDE project system
>> (global-ede-mode t)
>> ; minimal + custom (only activate what we need and use)
>> (global-semantic-idle-scheduler-mode 1)
>> (global-semantic-idle-summary-mode 1)
>> (global-semantic-idle-tag-highlight-mode 0)
>> (global-senator-minor-mode 1)
>> (global-semantic-decoration-mode 1)
>> (require 'semantic-decorate-include)
>> (global-semantic-show-unmatched-syntax-mode 1)
>> (if (fboundp #'which-func-mode)
>> (add-hook 'semantic-init-hooks (lambda ()
>> (which-func-mode 1))))
>> ; Options
>> (global-semantic-tag-folding-mode 1)
>> ; useful decoration for tags / symbols
>> (require 'semantic-ia)
>> ; highlight compilation errors
>> (require 'lmcompile)
>> (add-hook 'compilation-finish-functions 'vj-compilation-finish-highlight)
>> (defun vj-compilation-finish-highlight (buffer result-str)
>> ; ask gcc for config for cedet
>> (require 'semantic-gcc)
>> (semantic-gcc-setup "gcc")
>> ; semanticdb
>> (require 'semanticdb)
>> (global-semanticdb-minor-mode 1)
>> ; semanticdb + global
>> (require 'semanticdb-global)
>> (semanticdb-enable-gnu-global-databases 'c-mode)
>> (semanticdb-enable-gnu-global-databases 'c++-mode)
>> ; srecode
>> (require 'srecode)
>> (global-srecode-minor-mode 1)
>> ;;--------------------- End of Load of CEDET
>> Eric M. Ludlam wrote:
>>>>>> Damien Deville <damien.deville@...> seems to think that:
>>>> Since a few days the two following modes are buggy
>>>> Here is what i see:
>>>> - open a foo.c file where function foo() is defined as a static function
>>>> - navigate to any place where foo is called
>>>> - point the cursor to the start of foo();
>>>> - idle for a while
>>>> - any of the two previously mention idle mode will fire and the buffer
>>>> foo.c and *scratch* is displayed
>>>> - buffer *Messages* contains
>>>> Idle Service Error semantic-idle-summary-idle-function: "#<buffer
>>>> *scratch*> - Wrong type argument: syntax-table-p, nil"
>>> [ ... ]
>>> Hi again.
>>> I changed my mind. I saw that in the idle timer it said (roughly)
>>> (dolist (a b)
>>> ;; call an idle function
>>> so I changed it to:
>>> (dolist (a b)
>>> ;; Call an idle function
>>> and I'll take a wild guess that should fix this issue.
>>> I checked it into CVS. Let me know how it goes.
NETASQ - We secure IT