[omniORB] LongLongSeq undefined

[Thomas BARON]

Hi,

Well, I tried yesterday the latest snapshot from CVS, and the build were
successfull, no problem at all, thank you :o) Now I'm waiting for the 4.0.1 !
About the 4.0.0 release, I delete my source directory to untar again the tarball
and same problem : the configure script still find that 'long long' type is 8
bytes. So it's not a problem coming from a cached result from another platform.
I've not try the 'omniidl -v -P' command yet, I'll try later and tell you about
the output...

Regards,

Tom

Duncan Grisby a écrit :

> On Tuesday 28 January, Thomas BARON wrote:
>
> > I didn't check if 'long long' type exists on IRIX by myself, but the
> > configure script outputs that 'long long' type exists, with a size of 8
> > bytes. And the only options I gave it was about C and C++ compilers,
> > Python path (version 2.2.1), and the target path for installation.
> > Has_LONGLONG is defined in include/omniORB4/CORBA_sysdep_auto.h, so I
> > don't warn myself about dynAny believing that 'long long' type exists.
> > But in the build process, I saw nothing about a compilation by omniidl of
> > the .idl file (I don't remember the file name, sorry) in which LongLongSeq
> > is defined. So maybe something like that is missing ?
>
> LongLongSeq is defined in corbaidl.idl, which is compiled during the
> omniORB build. Please can you run this command
>
>   omniidl -v -P foo.idl
>
> giving it any IDL file. That will print the command given to the
> pre-processor. It should include a definition of -DHAS_LongLong (and
> -DHAS_LongDouble if you have it).
>
> I don't suppose you have installed the output from a build on a
> different architecture in the same place?  That could have overwritten
> the version of corbaidl_defs.hh with one that doesn't include the long
> long sequence.
>
> Cheers,
>
> Duncan.
>
> --
>  -- Duncan Grisby         --
>   -- duncan@grisby.org     --
>    -- http://www.grisby.org --