On date Monday 2008-02-11 15:19:55 -0500, Eric M. Ludlam wrote:
> >>> Stefano Sabatini <stefano.sabatini-lala@...> seems to think that:
> >On date Monday 2008-02-11 07:39:42 -0500, Eric M. Ludlam wrote:
> >> Hi,
> >>
> >> I don't know where that message might come from. Is speedbar or ECB
> >> active at the time? I posted an ECB patch a few days ago to make it
> >> compatible w/ CEDET/CVS.
> >>
> >> If not, try:
> >>
> >> M-x toggle-debug-on-error RET
> >> and see if it will spawn a stack-trace.
> >
> >I deactivated every call to ECB in my init.el (cedet should work
> >anyway, right?), but I still continue to get those strange
> >errors. Nontheless, when I edebug-eval-defun I discovered that
> >semanticdb-load-enable-minimum-features seems to be executed
> >correctly, and its evaluation finally returns nil, so maybe I can skip
> >this problem for now.
> >
> >Strange enough the toggle-debug-on-error didn't managed to raise any
> >stack-trace.
>
>
> The error is likely coming out of an idle timer or some-such which
> disables the debugger.
>
>
> >> If you were using an older version of CEDET before, it might be
> >> worth deleting your ~/.semanticdb directory, or any semantic.cache
> >> files you have lying about.
> >
> >Good catch, I tried with:
> >find . -name 'semantic.cache' | xargs rm
> >
> >but the error messages seem to stay.
> >
> >Anyway now I have my file displayed in the active buffer, and the
> >semantic-load-enable-minimum-features has been likely loaded. Now what
> >should I expect?
> >
> >I tried to manually launch semantic-idle-summary-mode and I got the message:
> >semantic-idle-summary-mode-setup: Buffer DumbThing.cpp was not set up for parsing
> >
> >(DumbThing.cpp is the name of the test file I'm using).
> >
> >Finally I tried with semanticdb-dump-all-table-summary and I can't see
> >no sign of the test file.
> >
> >What am I missing?
> [ ... ]
>
> You seem to be having a tough time of it. This error sounds like the
> autoloads file wasn't built properly. The setup for C++ mode is
> handled in a hook which is run via the autoloads file, which is loaded
> by the cedet.el script.
>
> Re-issuing a "make" from the cedet directory should clear and rebuild
> the autoload files. A "make bootstrap" will clean up even more stuff
> and try again.
>
> Another option would be to just delete everything, and start from
> scratch.
>
>
> The CEDET install should be as simple as:
> 1) build
> 2) Install a couple lines in your .emacs for load, and semantic
> 3) Use the tool.
>
> See the steps in the cedet/INSTALL file for more.
>
> Good Luck
Hi Eric, you was right, a simple make bootstrap seemed to fix all the
problems, now I can see *for the first time* the autocompletion
mechanism at work, OOORAY!! :-)
Many thanks!!
Best regards.
--
Stefano Sabatini
Linux user number 337176 (see http://counter.li.org)
|