cdt indexer uses gcc ? [message #183422] |
Wed, 31 January 2007 07:20  |
Eclipse User |
|
|
|
Originally posted by: asdf.dd.dd
Hello,
As I understand it , gcc is launched to generate some of the information
requred to make the code assis / go-to-definition logic work.
Is that that only true for full indexer , or also for fast indexer ?
I have a standard make project where the indexing works terrible.
The compiler used is a special TI compiler for that paticular embedded chip.
So my question is, will the indexing only work satisfactory if the source
code can more or less compile in the gcc compiler ?
Maybe a good advice would be, that if you don't use the gcc compiler for the
actual compiling, dont expect the full indexer to work without problems ?
Can I launch the "gcc scanner command line" my self , and examie the output
files, to slowly find out why this project is such a challenge for the cdt
indexer?
(as mentione in other post, i am currently evaluating the 4m4 also, it
allready looks much better.)
thx for any advice
Troels
|
|
|
|
Re: cdt indexer uses gcc ? [message #183565 is a reply to message #183422] |
Thu, 01 February 2007 07:01  |
Eclipse User |
|
|
|
Troels,
CDT's built-in parser (which in turn feeds the indexer) needs to know the
built-in paths of your compiler. This is achieved with the scanner discovery
service. You can set some of it's parameters in the project properties,
C/C++ Make Project, Discovery Options.
You may try the profiles which are included in CDT (gcc per project and per
file) and you will probably want to change the compiler command so that your
(special, cross) compiler is called.
Depending on how gcc-like your compiler is the CDT will be able to make more
or less sense of it's output.
If it is not very gcc-like then you may need to create your own scanner
discovery profile by supplying a plug-in.
HTH
Norbert
|
|
|
Powered by
FUDForum. Page generated in 0.09107 seconds