[GCC-XML]Status code 33

Jacek Generowicz jacek.generowicz at cern.ch
Fri May 30 04:34:55 EDT 2003


Brad King <brad.king at kitware.com> writes:

> Okay, I just looked back into the front-end code, and I see that when the
> patched GCC parser is run internally, it uses that program's result code.
> Therefore, result code 33 is coming from GCC.

OK, thanks.

I didn't pursue this to its conclusion, as I found out what was going
wrong: the include path was incorrect, so none of the headers being
included were actually found. However, gccxml reports "Success" for
each of them.

I've tried to reproduce the behaviour in a minimal set of files:

   == /path/FOO/foo.h: ========
   class foo {
     int bar;
   };
   == /path/FOO/outer.h =======
   #include "FOO/foo.h"
   ============================

Now, the command

   gccxml -I "/path/FOO" /path/FOO/outer.h -fxml=outer.xml

sends the following to stderr

   /afs/cern.ch/user/j/jacek/public/seal/Scripting/WrapperEvaluation/src/FOO/outer.h:1:21: FOO/foo.h: Success

and leaves the following in outer.xml

   <?xml version="1.0"?>
   <GCC_XML>
     <Namespace id="_1" name="::" members=""/>
     <Namespace id="_2" name="std" context="_1" members=""/>
   </GCC_XML>

while removing the "FOO" from the include-path in the command, produces the expected result:

   <?xml version="1.0"?>
   <GCC_XML>
     <Namespace id="_1" name="::" members="_3 "/>
     <Namespace id="_2" name="std" context="_1" members=""/>
     <Class id="_3" name="foo" context="_1" location="f0:1" members="_4 _5 _6 " bases=""/>
     <Field id="_4" name="bar" access="private" type="_7" context="_3" location="f0:2"/>
     <Constructor id="_5" name="foo" context="_3" location="f0:1">
       <Argument name="_ctor_arg" type="_8"/>
     </Constructor>
     <Constructor id="_6" name="foo" context="_3" location="f0:1"/>
     <FundamentalType id="_7" name="int"/>
     <ReferenceType id="_8" type="_3c"/>
     <File id="f0" name="/path/FOO/foo.h"/>
   </GCC_XML>


Bug or Feature ?

(Maybe just some meaning of the word "Success" with which I am not
familiar :-)



More information about the gccxml mailing list