Re: [CEDET-devel] Configuring Cedet for large projects? (Apologies if this is the wrong mailing lis
Brought to you by:
zappo
From: Eric M. L. <er...@si...> - 2010-04-19 14:20:31
|
You may need to do a clean rebuild. The 'interactively' changes appears to be a bit finicky to update to. Clear rebuilds have helped others get around this. I'm not 100% sure about the cscope angle though. It is working for me though. Eric On 04/19/2010 08:38 AM, Granville Barnett wrote: > Sorry for the slight delay. I downloaded and built HEAD (#8250) and now > I get (using the same configuration as previous) the following error: > > Invalid function: cedet-called-interactively-p > > When starting emacs with --debug-init it gives the trace attached. > > The error seems to be something to do with checking cscope, the last > function call before the error function is > cedet-gnu-global-version-check. If I disable the gnu-global modes for c > and c++ modes then the error no longer occurs. > > Thanks, > > Granville > > PS. I've just noticed I was emailing you direct, I've added the mailing > list now for others that may come across the same problem. Apologies. > > On 18 April 2010 14:21, Eric M. Ludlam <er...@si... > <mailto:er...@si...>> wrote: > > Hi, > > Thanks for the info. > > Your specific example didn't repro for me because I think I have a > different version of Linux source. I did find a different symbol > that showed that kind of problem, however, and I was able to "fix" > it in semanticdb-typecache.el. > > The issues was that something was causing the current buffer to > change, and that was throwing everything off. Hopefully this will > fix it for you. If not, let me know. > > Lastly, my linux specific EDE project should still work for you, as > it was working fine for me, even in reproducing the problem you > described. If you disable the ede-cpp-root project you created, and > then do: > > M-x semantic-c-describe-environment > > does it show ede-linux as the current project? > > Eric > > > On 04/13/2010 07:46 AM, Granville Barnett wrote: > > Hi Eric, > > Thanks for the quick response. > > I had to setup an EDE project for Linux source as I had a tonne of > unparsed includes all the time, when I set the EDE project up they > seemed to be mostly resolved (I can't vouch for all files as I don't > read them all). > > To reproduce (and I'm using a random example) open > kernel/mutex.c and go > to line 133 and do a quick jump on the symbol > __mutex_fastpath_unlock > and you will get the error I mentioned assuming you have created a > Global database for the source code. > > FYI (and this may be wrong) my ede configuration is as follows: > > (ede-cpp-root-project "Linux Kernel" > :name "Linux Kernel" > :file "~/git/code/linux-2.6.33.1/README" > :include-path '("/" > "/include")) > > I'm using the following settings for Cedet: > > ;; cedet > ;; > (load-file "~/.emacs.d/vendor/cedet-1.0pre7/common/cedet.el") > (semantic-load-enable-excessive-code-helpers) > (global-ede-mode t) > (require 'semantic-ia) > (require 'semantic-gcc) > (require 'eassist) > > ;; gnu global support > (require 'semanticdb-global) > (semanticdb-enable-gnu-global-databases 'c-mode) > (semanticdb-enable-gnu-global-databases 'c++-mode) > > (require 'semantic-decorate-include) > > Thank you for clarifying what the error means. > > Unfortunately I'm not familiar with edebug. > > Thanks for the help. > > Granville > > > On 13 April 2010 12:27, Eric M. Ludlam <er...@si... > <mailto:er...@si...> > <mailto:er...@si... <mailto:er...@si...>>> > wrote: > > Hi, > > CEDET comes with a built-in EDE project for Linux, so you > should not > need to configure your own EDE project. I'm curious about any > feedback people have about this custom EDE project type so > it can be > made better, as I am not a Linux kernel hacker, and guessed > about > some things. > > If you can provide a specific example; ie - a file, line and > symbol > where you see this issue, I can probably reproduce it here in my > Linux sources (as installed by Ubuntu) to see what's going on. > > As it stands, the error with the text "no buffer > information" means > that a tag was found, but it is not associated with a > buffer. I can > imagine that this is related to some GNU Global or other > symbol that > didn't get resolved down to a buffer location for some reason. > > If you are familiar with running edebug, you can use > edebug-defun on > semantic-ia--fast-jump-helper, and see what's going on. > > Good Luck > Eric > > > On 04/13/2010 07:08 AM, Granville Barnett wrote: > > Hi, > > I'm trying to get Cedet 1 pre7 to work well while I'm > hacking on the > Linux kernel - my results are somewhat erratic at the > moment, > however. > > I have done the following: > > - Setup Cedet to use GNU Global > - Created Global databases for the kernel source code > (I've also > generated databases for /usr/include and /usr/local/include) > - Created an Ede project listing a few include directories > > (I've used some of the configuration settings from > http://alexott.net/en/writings/emacs-devenv/EmacsCedet.html as > the basis > for my Cedet configuration.) > > Now when I go to jump to definitions (using fast jump) I > sometimes jump > to the correct location but other times I just get a > message > stating > that no buffer information exists for a particular > symbol. I'm > not a pro > Cedet user, in fact I've only recently started to use > it, but I > thought > that when one delegated symbol information to Global then it > should use > Global for jumping so I'm not entirely sure what this > message means. > > Also, if anyone has any guidance on how to best use > Cedet with large > codebases I would very much like to hear from them. > > Any help would be most appreciated. > > Kind regards, > > Granville > > > |