Re: [cdt-dev] Interesting search/indexing results

[John Camelon <jcamelon@xxxxxxxxxx>]

This is why we need a target model post-1.2, because otherwise the user 
shall get heavily involved in 
concepts that should be kept hidden to them. 

The scanner handles preprocessor errors the way a preprocessor would. This 
could change once we 
implement a full IProblem error handling strategy.  The rationale is that 
if we were not able to include something
or our macros are set improperly, we will encounter so many errors down 
the road that the parse very well
could end up useless. 

JohnC

cdt-dev-admin@xxxxxxxxxxx wrote on 10/10/2003 08:40:43 AM:

> Hi Thomas,
> Some people have taken a longer long weekend, so its a little quiet 
here. 
> But I can answer your first question.
> Yes, a ScannerException is thrown because of the #error, and yes it will 

> stop the parse for that compilation unit which will cause that file to 
not 
> be completely indexed.
> 
> To get this to work you will need to add one of the compiler symbols ( 
> __MWERKS__,  __GNUC__, etc) to the symbols section of the build 
settings. 
> And also the path the those include files.  I'm assuming that _NTO_HDR_ 
is 
> defined somewhere.
> 
> -Andrew
> 
> 
> 
> Thomas Fletcher <ThomasF@xxxxxxx> 
> Sent by: cdt-dev-admin@xxxxxxxxxxx
> 10/10/2003 08:43 AM
> Please respond to
> cdt-dev@xxxxxxxxxxx
> 
> 
> To
> "'cdt-dev@xxxxxxxxxxx'" <cdt-dev@xxxxxxxxxxx>
> cc
> 
> Subject
> [cdt-dev] Interesting search/indexing results
> 
> 
> 
> 
> 
> 
> Hey folks, (we are having a use fest today, so expect a lot of traffic)
> 
>   So I was spending some time trying to make the world a better place
> yesterday by adding a few "convenience" extensions into the C/C++ Editor
> and kind of hit a snag (or two).  These haven't been bugzilla'ed yet
> since I wanted feedback on if they were in fact errors:
> 
> 1) In a QNX C Project (a normal C project but with QNX extensions for
> our environment) I don't get any indexing being performed. It used to
> work (before we actually included the system headers) so I took a look
> at the log and I saw:
> 
> ENTRY org.eclipse.cdt.core {DATE & TIME}
> MESSAGE Scanner Exception: #error not configured for compiler
> 
> This is likely due to the behaviour in the QNX Neutrino header files
> which does things like:
> 
> #if defined(__MWERKS__)
> #include _NTO_HDR_(sys/compiler_mwerks.h)
> #elif defined(__WATCOMC__)
> #include _NTO_HDR_(sys/compiler_watcom.h)
> #elif defined(__GNUC__)
> #include _NTO_HDR_(sys/compiler_gnu.h)
> #elif defined(__HIGHC__)
> #include _NTO_HDR_(sys/compiler_highc.h)
> #else
> #error not configured for compiler
> #endif
> 
> Is the exception thrown (if it is really an exception and not just a log
> message), causing the indexing to stop?
> 
> because ....
> 
> 2) The simple project that I have looks like:
> 
> --- file1.c ---
> #include <stdlib.h>
> #include <stdio.h>
> 
> int main(int argc, char **argv) {
>                  printf("Hello CDT folks \n");
>                  afunction(argc, argv);
>                  return EXIT_SUCCESS
> }
> 
> --- file2.c ---
> #include <stdlib.h>
> #include <stdio.h>
> 
> int afunction(int argc, char **argv) {
>                  printf("I'm lazy so I cut and paste\n");
>                  return EXIT_SUCCESS
> }
> 
> ---
> 
> This code doesn't get indexed in the qnx project, but in the standard 
make
> it does.  So then using the search I do a search for afunc* which 
returns
> only the instance in file2.c, even when I specified All Occurrences. How
> come the reference to the function in main() is not returned?
> 
> 3) Cacheing of the index file(s)?
> 
> Is this being done currently?  It seems to be iterating/indexing lots. 
No
> specifics on this just yet.
> 
> Thanks guys,
>  Thomas
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/cdt-dev
> 
> 
> _______________________________________________
> cdt-dev mailing list
> cdt-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/cdt-dev