CEDET

Hi,

CEDET was certainly designed to work cross-language, as it provides a 
single API that works across all languages, with a secondary API for 
working across languages that have basic structure, like javascript or C.

Unfortunately, I'm not a Java or JDEE expert.

The basic strategy is to get an EDE project pointing to the root of all 
your code.  I'm assuming all the languages you mentioned are in one big 
mega-project.  That top level project then needs to know about all the 
targets.  For simplicity sake, you could have one target per language 
type, or more detailed, one target per language/directory.

That sounds good, but no-one has written such a generic project.  If you 
feel adventurous, you could copy ede-emacs.el to your own thing, and 
change ede-emacs-project-root to ede-myproject-project-root and fix the 
directory detection.  The framework for rolling your own is pretty easy 
from there.

If all else fails, you can just use ede-cpp-root, and it will deal with 
your c++ code fine, and provide a simple root-of-project for all the 
cross referencing functions.

As for all your languages, each language has varying levels of support 
based on how much effort the folks who use that language have chosen to 
put into it.  In the JDEE case, they have their own completion system. 
There have been projects to merge CEDET's from-source-only completion 
with JDEEs from-compiled-only parsing, but it has never been finished.

As for the symbol references, you can use any of the different tools you 
prefer, or none.  If you try it (once you have an EDE project specifying 
the root) with the find/grep solution and it is fast enough, then you 
can consider yourself done.

The simple-recipe version of the doc is in cedet/common/cedet.el where 
it will help based on what kind of task you have.  Your task isn't in 
there, but some aspects of it are.

If you have more specific questions, they may be easier to answer.

Eric

On 04/30/2010 03:40 PM, David Ventimiglia wrote:
> Hello,
>
> I’m sure versions of this newbie type question get asked again and again
> from time to time, but I just couldn’t pull together an answer for
> myself by sifting the CEDET list archives, so (apologies), I’ll ask away.
>
> What are some best practices for configuring CEDET for a Java-ish
> project, in order to achieve:
>
> * Smart completion
>
> * Symbol lookup
>
> * File name lookup
>
> * Cross-references
>
> I should point out that my "project" (as is quite common) has many file
> types besides Java: XML, HTML, JavaScript, some C/C++ code, etc., and
> I'd like to have a somewhat uniform Emacs view of all this. That's
> partly why JDEE project files don't quite do the trick.
>
> The problem I'm having is that the JDEE provides some version of these
> features, with some deficiencies, and seemingly with considerable
> overlap with CEDET. Completion and symbol lookup exist, but are slow and
> only work for compiled classes. Also, they only work for Java files,
> which is inadequate when a project has a mixture of other programming
> languages and file types. File name lookup seems not to exist at all.
> And JDEE has the xref database, but it's a little flaky (it pukes while
> parsing one of my files) and again, only works for Java files.
>
> So, I'd prefer to use CEDET's features for these. Evidently, it involves
> some combination of CEDET customizations, EDE project configuration, and
> maybe some use of idutils and/or GNU Global for good measure.
> Unfortunately, I haven't yet figured out how to do this, despite staring
> at the CEDET documentation for a few days. I'll keep plugging away at
> it, but I confess I'm badly in need of a clue on this. Thanks in
> advance. And I should say, I really am very impressed by CEDET (despite
> an imperfect understanding of it) and sense considerable power embedded
> in it.
>
> All the best,
>
> David Ventimiglia
>
>
>
> ------------------------------------------------------------------------------
>
>
>
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic

Thanks, Eric.  I've a few, more-specific questions spliced in below.

> The basic strategy is to get an EDE project pointing to the root of
all
> your code.  I'm assuming all the languages you mentioned are in one
big
> mega-project.  That top level project then needs to know about all the
> targets.  For simplicity sake, you could have one target per language
> type, or more detailed, one target per language/directory.

So far, I just started with ede-simple-project, which I created with the
command ede-new while visiting a Java file buried deep in that part of
our source-tree.  This put a file in my $HOME/.ede directory defining
that ede-simple-project.  I also have run GNU Global in my project root.
What do I do now to enable completion, lookup, and cross-references for
symbols?  

> As for the symbol references, you can use any of the different tools
> you
> prefer, or none.  If you try it (once you have an EDE project
> specifying
> the root) with the find/grep solution and it is fast enough, then you
> can consider yourself done.

I made these customizations:

(require 'cedet)
(global-ede-mode 1)
(semanticdb-enable-gnu-global-databases 'java-mode)
(semanticdb-enable-gnu-global-databases 'jde-mode)
(setq ede-locate-setup-options
      '(ede-locate-global
        ede-locate-base))

But when I invoke the command semantic-symref-symbol on some symbol in a
Java file, I get the message:

semantic-symref-tool-grep - Needs to be configured for jde-mode

What does this mean?

Finally, must one use Global, idutils, locate, etc., to do symbol
lookups?  Is it possible just to rely on the CEDET tools alone?  What I
mean is, I gather that semantic is parsing and caching the tags for,
say, the Java files that I visit.  Is it possible to have semantic parse
ALL my Java source files ahead of time (a la Global, idutils, ctags,
etc.), then just use semanticdb to support completion, lookup, and
cross-referencing?  This isn't a huge deal, but I'm trying to evangelize
Emacs+CEDET+JDEE at my company, which is a Windows shop, and the more
external tools I have to integrate (like finding Windows versions of
Global, idutils, etc.), the harder that sell becomes.  Or, would it
simply be too much of a performance hit to ask semantic to do what
Global et al are optimized to do very fast?

Thanks!
Best,
David

Hi,

I have not used the Emacs integrated version of CEDET.  I do know that 
the setup isn't the same.  You have to turn on 'semantic-mode', or 
something like that.  I don't know what the state of the doc is for it.

Basic customizations (variables, modes, etc) should be the same, but 
without semantic turned on nothing will work because the modes disable 
themselves in buffers that aren't supported by a parser.

Eric

On 04/15/2010 01:28 PM, Suvayu Ali wrote:
> Hi everyone,
>
> I have been using CEDET for quite a while now, it works great with the
> cvs version checked out and compiled in my ~/.emacs.d/elisp. However
> lately I have been trying to setup the version of CEDET that comes
> with the latest Emacs 23.1.95. No matter what I do, I can't seem to
> get any of my customisations working. (they are quite simple, setting
> up some include paths and some useful key bindings)
>
> Upon doing `M-x load-library RET cedet', `M-x load-library RET ede'
> and `M-x load-library RET semantic' and doing `M-x cedet-version' I
> get the following.
>
> CEDET Version:  1.0pre7
>                           Requested       File            Loaded
>     Package               Version         Version         Version
>     ----------------------------------------------------------
>     cedet:                1.0pre7         nil             ok
>     eieio:                1.2             nil             ok
>     semantic:             2.0pre7         nil             ok
>     srecode:              1.0pre7         nil             Not Loaded
>     ede:                  1.0pre7         nil             ok
>     speedbar:             1.0.3           nil             1.0
>
> Where as with the cvs version I get something like this,
>
> CEDET Version:	1.0pre7
>     			Requested	File		Loaded
>     Package		Version		Version		Version
>     ----------------------------------------------------------
>     cedet:		1.0pre7		ok		ok
>     eieio:		1.2		ok		ok
>     semantic:		2.0pre7		ok		ok
>     srecode:		1.0pre7		ok		ok
>     ede:			1.0pre7		ok		ok
>     speedbar:		1.0.3		ok		ok
>     cogre:		1.0pre7		ok		Not Loaded
>     cedet-contrib:	1.0pre7		ok		Not Loaded
>
> Any pointers about what I am doing wrong?
>

Hi Eric,

On 18 April 2010 05:04, Eric M. Ludlam <eric@...> wrote:
> Hi,
>
> I have not used the Emacs integrated version of CEDET.  I do know that the
> setup isn't the same.  You have to turn on 'semantic-mode', or something
> like that.  I don't know what the state of the doc is for it.
>
> Basic customizations (variables, modes, etc) should be the same, but without
> semantic turned on nothing will work because the modes disable themselves in
> buffers that aren't supported by a parser.
>

Many functions and variable names seem to be different in the CEDET
that comes with Emacs 23.1.95. For example I couldn't find easisst. So
I finally gave up and thinking of installing the CVS version. I 'll
try my hand on this again when I get the time. If I get it working,
I'll keep post back to the list. And thanks for your reply Eric.

> Eric
>

-- 
Suvayu

Open source is the future. It sets us free.

On 04/21/2010 03:51 PM, Per Nordlöw wrote:
> If we choose the GNU GLOBAL backend for Symref, I believe it should
> automatically on-demand create (gtags) and update (global -u) a GNU
> GLOBAL database in current project root. We could use advices for that
> but it would be cleaner if we add a hook, say (defun symref-pre-hook
> (backend directory)), in some suitable place.

Hi,

The simplest method for you would be to use CLOS to do what you want as 
a one off configuration.  For example:

(defmethod semantic-symref-perform-search :BEFORE ((tool 
semantic-symref-tool-global))
    "Method called before the symref search is performed."
     (cedet-global-update))

though someone needs to write 'cedet-global-update'.

As a general feature, this is a pretty good idea.

Updating tags tables seems like a feature for EDE to handle, as I assume 
gtags databases are related to projects.  Symref could then dispatch to 
EDE, which it already does in semantic-symref-calculate-rootdir.

When contemplating where to put "update tags", the "Project" menu of EDE 
seems logical, which also indicates that EDE may be a good place for it.

I've also never gotten ETAGS support in.  ETAGS can do some things that 
GLOBAL and CSCOPE do, just slower, and ETAGS is something that shows up 
in a lot of Makefiles with a "tags" target, making it a good candidate 
for EDE support in this concept also.

Stuff to think about.
Eric

On 04/14/2010 07:03 AM, Per Nordlöw wrote:
> Does Semantic support some kind of completion of arguments when writing
> function calls?

The core engine for doing completion knows what argument you are in for 
a particular call, and makes decisions based on that when providing 
completions.

It doesn't make assumptions, such as you are more likely to use some 
local variable that misc other options.

> I believe this is one of the most important features of any IDE and my
> guess is that it shouldn't be too complicated to implement it CEDET.

Your description so far is fairly vague, so I'm not sure what is not 
complicated yet.

> We could probably "borrow" code from template/snippet libraries, for
> example yasnippet, for the argument navigation/modification part.
> I like its in-place-navigation interface.

If yasnippet has been assigned to the FSF, then it would make a fine 
baseline for UI based work with CEDET.  I have not put much effort into 
the completion UIs in CEDET as company-mode, and competion-ui both are 
nicer, and now integrate with CEDET's smart completion engine.  There 
are a lot of nice UIs out there for this sort of thing.

As for the in-place field editing, there is a primitive variant I wrote 
in SRecode for that sort of thing,  but the yasnippet one is nicer.  It 
would be great to use yasnippet's if it is available, as I haven't had 
time to fully implement my field-editor.

As for replacing srecode with yasnippet completely, the yasnippet 
template system is not powerful enough for what I need to do with 
automated code generation, or creating generic code generating 
applications that are language agnostic.

> And the function signatures already exists in Semantic right?

Yup.

> What semantic-function do I use to iterate over these function arguments?

If you write code like this:

(let ((ctxt (semantic-analyze-current-ctxt)))
    ...)

while point is in code like this:

int main(int argc, char*argv[]) {
   main(1,!);  // Cursor is the !
}

Then the ctxt has slots like this:

  ] :function #<TAG LIST: 1 entries>
    * main : int
  ] :index 2
  ] :argument #<TAG LIST: 1 entries>
    * argv : char


which you can examine in detail with the command semantic-adebug-analyze.


If you describe the workflow you are thinking of, I might be able to 
provide a more targeted answer.

Eric

On 04/08/2010 06:26 PM, Wei Li wrote:
> Hi all,
>
> I am trying to setup CEDET's smart completion. However, when I try to
> bind the semantic-ia-complete-symbol-menu and use it, it shows "Wrong
> type argument: syntax-table-p, nil". How is this function working? I
> thought it means that if I type partial name of e.g. a function and
> trigger this binding it will show the matching items in a menu (and I
> remember I saw such functionality on somebody's machine). I tried search
> the web but got no luck. Is there anything else I need to set? Here is
> my command:
>
> (define-key c-mode-base-map [(meta ?/)] 'semantic-ia-complete-symbol-menu)
>
> And of course, the above command is within my hook definition added to
> c-mode-common-hook. Any help will be appreciated.

You certainly have the right idea, but are running into some other kind 
of problem.  What version of CEDET are you using?

If you enable:

M-x toggle-debug-on-error RET

it will produce a stack trace which will help reveal what is going on.

Eric

On Fri, Apr 9, 2010 at 3:17 AM, Eric M. Ludlam <eric@...:

> On 04/08/2010 06:26 PM, Wei Li wrote:
>
>> Hi all,
>>
>> I am trying to setup CEDET's smart completion. However, when I try to
>> bind the semantic-ia-complete-symbol-menu and use it, it shows "Wrong
>> type argument: syntax-table-p, nil". How is this function working? I
>> thought it means that if I type partial name of e.g. a function and
>> trigger this binding it will show the matching items in a menu (and I
>> remember I saw such functionality on somebody's machine). I tried search
>> the web but got no luck. Is there anything else I need to set? Here is
>> my command:
>>
>> (define-key c-mode-base-map [(meta ?/)] 'semantic-ia-complete-symbol-menu)
>>
>> And of course, the above command is within my hook definition added to
>> c-mode-common-hook. Any help will be appreciated.
>>
>
> You certainly have the right idea, but are running into some other kind of
> problem.  What version of CEDET are you using?
>
> If you enable:
>
> M-x toggle-debug-on-error RET
>
> it will produce a stack trace which will help reveal what is going on.
>
> Eric
>

Hi Eric,

Thanks for your quick answer, even to my kinda fool question. I googled
using the debug information and got a post of a similar problem, and you
replied there saying it might be because senator was not on. And I found a
problem: In Alex Ott's tutorial he said for semantic's customization, users
need to choose one command from semantic-load-enable-minimum-features,
semantic-load-enable-code-helpers,
semantic-load-enable-gaudy-code-helpers,
emantic-load-enable-excessive-code-helpers,
semantic-load-enable-semantic-debugging-helpers, and each command contains
all the features of previous commands. However in another tutorial the
author used two commands, e.g semantic-load-enable-code-helpers and
semantic-load-enable-semantic-debugging-helpers.  I didn't have Senator on
using Alex's setting (with only semantic-load-enable-semantic-debugging-helpers
command) but had it using the other setting. Does it mean that
semantic-load-enable-semantic-debugging-helpers
doesn't enable Senator?

BTW, I used 1.0pre6 when I asked in the first place and now I am using
1.0pre7. Thanks for your help!

Cheers

On 04/09/2010 06:49 PM, Wei Li wrote:
>
> Hi Eric,
>
> Thanks for your quick answer, even to my kinda fool question. I googled
> using the debug information and got a post of a similar problem, and you
> replied there saying it might be because senator was not on. And I found
> a problem: In Alex Ott's tutorial he said for semantic's customization,
> users need to choose one command from
> |semantic-load-enable-minimum-features,
> ||semantic-load-enable-code-helpers,
> ||semantic-load-enable-gaudy-code-helpers,
> ||emantic-load-enable-excessive-code-helpers,
> ||semantic-load-enable-semantic-debugging-helpers|, and each command
> contains all the features of previous commands. However in another
> tutorial the author used two commands, e.g
> |||semantic-load-enable-code-helpers |and
> |semantic-load-enable-semantic-debugging-helpers.|  I didn't have
> Senator on using Alex's setting (with only
> |semantic-load-enable-semantic-debugging-helpers |command) but had it
> using the other setting. Does it mean that
> |semantic-load-enable-semantic-debugging-helpers |doesn't enable Senator?


Thanks for the detailed description of what you found.  Knowing that it 
was related to no semantic special features being enabled, I was able to 
reproduce the issue, and make Semantic robust to this situation.  It 
should now be possible to use the basics of smart completion without any 
of the extra support features.  It just won't do much. ;)

I checked in my changes to CVS.

Eric

Hi,

The best place to start is with the command 
semantic-c-describe-environment.  It may list some EDE object that will 
be doing the file finding for your subproject, depending on what it 
detected.

You can also right click on a header file, and select 'what is this' 
from the menu, and it might provide some (similar) hints.

In general, it is better to use an ede project type, even if it is 
ede-cpp-root, to define the base of your project, as it will do a better 
job of finding in-project header files.

It may also be that there is a bug with includes with directory paths in 
them for your case.

Eric

On 04/05/2010 03:05 PM, Christian Thaeter wrote:
> We've split our source directory into sub-parts (but not sub-projects).
> Cross-includes are deliberately prefixed with the directory there.
> This setup is like following:
> the project root is at
>    /home/ct/src/project
>
> then there is a ./src dir which contains the subdirs for each part
>    /home/ct/src/project/lib/
>    /home/ct/src/project/backend/
>    /home/ct/src/project/include/
>    ...
>
> semanticdb-project-roots is set to ("/home/ct/src/project/")
>
> now when I
>   #include "lib/something.h"
> semantic tells me this is a 'Unknown Include', just
>   #include "something.h"
> would work for semantic but is not the way its intended to be used and
> of course breaks the build setup.
>
> I suspect this is some bug in the way semanticdb looks up the includes,
> any idea how to shed some light on this?
>
> I'm using cedet-1.0pre7 on debian lenny with emacs32 from backports.org.
>
>
> 	Thanks
> 		Christian
>
>
>
> Btw: I tried ede but that has servere problems with our project setup,
> thats non-recursive building with one top-level Makefile.am which in
> turn includes Makefile.in from the sub dirs. Fixing this looks like a
> challenge to me, If you are interested I can provide more information :).
>
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>

Eric M. Ludlam wrote:
> Hi,
> 
> The best place to start is with the command
> semantic-c-describe-environment.  It may list some EDE object that will
> be doing the file finding for your subproject, depending on what it
> detected.
> 
> You can also right click on a header file, and select 'what is this'
> from the menu, and it might provide some (similar) hints.
> 
> In general, it is better to use an ede project type, even if it is
> ede-cpp-root, to define the base of your project, as it will do a better
> job of finding in-project header files.

that did the trick, thanks

	Christian


> 
> It may also be that there is a bug with includes with directory paths in
> them for your case.
> 
> Eric
> 
> On 04/05/2010 03:05 PM, Christian Thaeter wrote:
>> We've split our source directory into sub-parts (but not sub-projects).
>> Cross-includes are deliberately prefixed with the directory there.
>> This setup is like following:
>> the project root is at
>>    /home/ct/src/project
>>
>> then there is a ./src dir which contains the subdirs for each part
>>    /home/ct/src/project/lib/
>>    /home/ct/src/project/backend/
>>    /home/ct/src/project/include/
>>    ...
>>
>> semanticdb-project-roots is set to ("/home/ct/src/project/")
>>
>> now when I
>>   #include "lib/something.h"
>> semantic tells me this is a 'Unknown Include', just
>>   #include "something.h"
>> would work for semantic but is not the way its intended to be used and
>> of course breaks the build setup.
>>
>> I suspect this is some bug in the way semanticdb looks up the includes,
>> any idea how to shed some light on this?
>>
>> I'm using cedet-1.0pre7 on debian lenny with emacs32 from backports.org.
>>
>>
>>     Thanks
>>         Christian
>>
>>
>>
>> Btw: I tried ede but that has servere problems with our project setup,
>> thats non-recursive building with one top-level Makefile.am which in
>> turn includes Makefile.in from the sub dirs. Fixing this looks like a
>> challenge to me, If you are interested I can provide more information :).
>>
>> ------------------------------------------------------------------------------
>>
>> Download Intel® Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> _______________________________________________
>> cedet-semantic mailing list
>> cedet-semantic@...
>> https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>>

On 04/05/2010 10:14 AM, Tian Fang wrote:
> Hi,
>
> Is there a way to use GNU GLOBAL database pre-generated for auto-complete?
>
> In the following example, global DB is pre-generated, which includes the
> structure defined in a.h. Then, in b.c, w/o relying on the "include
> a.h", I can still use the global db to auto-complete structure fields
> defined in a.h.
>
> <root>/a/a.h
> <root>/b/b.c
> <root>/GLOBAL_DB

The Emacs code that comes with GNU Global can do completions w/out 
relying on which includes are in a given file.  Semantic can also 
complete like that, but only on global symbols.  It cannot do smart 
completion, and derive datatypes of variables to know they are 
structures, without relying on header files.

If you have GNU Global, Semantic can look up some global symbols using 
it.  That helps when you are starting in a fresh directory where Emacs 
hasn't had a chance to parse all your files yet.

Eric

On Tue, Apr 6, 2010 at 10:11 PM, Eric M. Ludlam <eric@...:

> On 04/05/2010 10:14 AM, Tian Fang wrote:
>
>> Hi,
>>
>> Is there a way to use GNU GLOBAL database pre-generated for auto-complete?
>>
>> In the following example, global DB is pre-generated, which includes the
>> structure defined in a.h. Then, in b.c, w/o relying on the "include
>> a.h", I can still use the global db to auto-complete structure fields
>> defined in a.h.
>>
>> <root>/a/a.h
>> <root>/b/b.c
>> <root>/GLOBAL_DB
>>
>
> The Emacs code that comes with GNU Global can do completions w/out relying
> on which includes are in a given file.  Semantic can also complete like
> that, but only on global symbols.  It cannot do smart completion, and derive
> datatypes of variables to know they are structures, without relying on
> header files.
>

[Tian] Glad to hear that Semantic can use only global for completion, even
not a smart one. In this case, how should I config my .emacs? I already have
the following two enabled. But still get "unknown type" error msg when doing
sematic-ia-complete-symbol for a structure field.

(semanticdb-enable-gnu-global-databases 'c-mode)
(semanticdb-enable-gnu-global-databases 'c++-mode)

Thx,
Tian


> If you have GNU Global, Semantic can look up some global symbols using it.
>  That helps when you are starting in a fresh directory where Emacs hasn't
> had a chance to parse all your files yet.
>
> Eric
>

On 04/07/2010 09:32 AM, Tian Fang wrote:
>
>
> On Tue, Apr 6, 2010 at 10:11 PM, Eric M. Ludlam <eric@...
> <mailto:eric@...>> wrote:
>
>     On 04/05/2010 10:14 AM, Tian Fang wrote:
>
>         Hi,
>
>         Is there a way to use GNU GLOBAL database pre-generated for
>         auto-complete?
>
>         In the following example, global DB is pre-generated, which
>         includes the
>         structure defined in a.h. Then, in b.c, w/o relying on the "include
>         a.h", I can still use the global db to auto-complete structure
>         fields
>         defined in a.h.
>
>         <root>/a/a.h
>         <root>/b/b.c
>         <root>/GLOBAL_DB
>
>
>     The Emacs code that comes with GNU Global can do completions w/out
>     relying on which includes are in a given file.  Semantic can also
>     complete like that, but only on global symbols.  It cannot do smart
>     completion, and derive datatypes of variables to know they are
>     structures, without relying on header files.
>
>
> [Tian] Glad to hear that Semantic can use only global for completion,
> even not a smart one. In this case, how should I config my .emacs? I
> already have the following two enabled. But still get "unknown type"
> error msg when doing sematic-ia-complete-symbol for a structure field.
>
> (semanticdb-enable-gnu-global-databases 'c-mode)
> (semanticdb-enable-gnu-global-databases 'c++-mode)

[...]

>     If you have GNU Global, Semantic can look up some global symbols
>     using it.  That helps when you are starting in a fresh directory
>     where Emacs hasn't had a chance to parse all your files yet.

There are two kinds of searches.  There are searches that search an 
entire 'project' for a symbol.  That is when GNU Global will be used. 
This kind of search isn't used very often.

There are also more narrowly defined searches, that will only pull in 
the stuff you include.  The smart completion system only uses the more 
narrowly defined searches, so you need to have the structure in a header 
that is included.

To make sure GNU Global is properly enabled, you can run the command:

M-x semanticdb-find-test-translate-path

to dump out the databases being used.  Is GNU Global there, at the end 
of the list?)

It may be that GNU Global doesn't know about the struct either for some 
reason.  I haven't tried you explicit example yet.  I can give it a go, 
but I'm not sure when.

Eric

Hi,

I am not surprised.  The lexical step doesn't do string concatenation, 
and, in fact, the include part of the lexical step only accepts strings 
via regular expression.

If you look in semantic-c.el at semantic-lex-c-include-system, you will 
see the code that extracts the <stuff> from the buffer.  This would need 
to take an extra step to split the string apart looking for macros. 
That seems like a lot of overhead for an obscure feature, but you could 
try it and see if it has bad performance implications.

Eric

On 04/02/2010 02:24 AM, 钱锦锋 wrote:
>
> Example:
>
> A    - a dirctory
> a.h  - a file under the "A" directory
>   struct aa {
>   int i;
>   };
>
> main.c    - a file in the same level with the directory "A"
>   #define _X_ "A"   /*#define _X_ A don't work as well*/
>
>   #include <_X_/a.h>
>
> void main()
> {
>    aa test;
> }
>
> Problem:
> aa can't be found in the main.c file (use semantic-ia-fast-jump, and the
> include statement is marked as unkonwn include)
>
> Could anyone know how to fix this issue?
> Thanks
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
>
>
>
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic