[omniORB] Linux RPMs built

[Bastiaan Bakker]

Hi Thomas,

Thanks for making RPMs available. 

Duncan, can these RPMS specs get your blessing? It would be nice if the
spec file could be included in the distribution, as it would allow for
building RPMs with a simple 'rpmbuild -ta omniORB-4.x.y.tar.gz'.

A few random observations and questions:

* RPM rebuild on i386 RedHat Linux 7.3 succeeded after changing the
python version requirement from 2.0 to 1.5.2. Any particular reason you
required 2.0?

* RPM rebuild on i386 RedHat Linux 8.0 failed. Cause: declaration
conflict of function crypt() in /usr/include/unistd.h and
/usr/include/openssl/des.h. 
Actually the OpenSSL guys appear responsible, considering the relevant
odd bit in des.h:

#if !defined(PERL5) && !defined(__FreeBSD__) && !defined(NeXT)
char *crypt(const char *buf,const char *salt);
#endif
 
Thanks,

Bastiaan Bakker
LifeLine Networks bv


On Sun, 2002-10-06 at 01:50, Thomas Lockhart wrote:
> I've (finally) built RPMs for omniORB-4.0.0 and omniORBpy-2.0 and posted 
> them at
> 
>    http://www.fourpalms.org/pub/omniORB/Mandrake/9.0/RPMS/
> and
>    http://www.fourpalms.org/pub/omniORB/Mandrake/9.0/SRPMS/
> 
> They are built with ssl enabled and thread tracing disabled per Mike's 
> suggestion a few weeks ago. They have received minimal testing, but I 
> have been using similar RPMs built from a recent beta release for some 
> python work. I'll fix anything which seems broken so let me know if you 
> have trouble.
> 
> I had to (slightly) patch omniORBpy-2.0 to build against the installed 
> omniORB RPM. The patch file is included below and adds DESTDIR to the 
> installation directories and changes a reference from 
> OMNIORB_ROOT/idl/omniORB/ to OMNIORB_ROOT/share/idl/omniORB which seems 
> to be where omniORB itself wants stash its idl files.
> 
> I'm happy to post the RPMs to the official site or to have someone else 
> take over ownership if that would be better. In the meantime if someone 
> wants to build for other platforms I can post those too; I've got 
> Mandrake-8.0 and 8.1 RPMs building...
> 
> Oh, Mandrake-9.0 uses gcc-3.2 so congratulations on having code which 
> makes it through the compiler.
> 
>                     - Thomas
> ----
> 

> diff -cr omniORBpy-2.0.orig/mk/beforeauto.mk.in omniORBpy-2.0/mk/beforeauto.mk.in
> *** omniORBpy-2.0.orig/mk/beforeauto.mk.in	2002-09-06 21:34:26.000000000 +0000
> --- omniORBpy-2.0/mk/beforeauto.mk.in	2002-10-05 13:52:49.000000000 +0000
> ***************
> *** 43,49 ****
>   # Directories for installation
>   #
>   
> ! prefix        	 := @prefix@
>   exec_prefix   	 := @exec_prefix@
>   INSTALLTARGET 	 := 1
>   INSTALLINCDIR 	 := @includedir@
> --- 43,49 ----
>   # Directories for installation
>   #
>   
> ! prefix        	 := $(DESTDIR)@prefix@
>   exec_prefix   	 := @exec_prefix@
>   INSTALLTARGET 	 := 1
>   INSTALLINCDIR 	 := @includedir@
> ***************
> *** 275,281 ****
>   # files.
>   
>   vpath %.idl $(IMPORT_TREES:%=%/idl) \
> !             $(OMNIORB_ROOT)/idl/omniORB $(DATADIR)/idl/omniORB
>   
>   IMPORT_IDLFLAGS += -I. $(patsubst %,-I%,$(VPATH)) \
>   		       $(patsubst %,-I%/idl,$(IMPORT_TREES))
> --- 275,281 ----
>   # files.
>   
>   vpath %.idl $(IMPORT_TREES:%=%/idl) \
> !             $(OMNIORB_ROOT)/share/idl/omniORB $(DATADIR)/idl/omniORB
>   
>   IMPORT_IDLFLAGS += -I. $(patsubst %,-I%,$(VPATH)) \
>   		       $(patsubst %,-I%/idl,$(IMPORT_TREES))
> Only in omniORBpy-2.0/mk: beforeauto.mk.in.orig