From: Geoffrey F. <fu...@ga...> - 2002-01-22 19:44:58
|
Joao Cardoso writes: > On Tuesday 22 January 2002 8:38 am, Alan W. Irwin wrote: > > Grab bag of topics here. The 3rd one is the nasty bug. > > > > (1) Maurice brought up turning on dynamic drivers by default earlier. > > Since nobody has objected, and I think this will be of some benefit to our > > users, I have just committed the change. > > Shouldn't "configure" disable dyndrivers if we don't know how to make shared > libraries on that system? Otherwise we get errors. Great/important point. I totally missed that. I just configured cvs head on Solaris (which worked fine yesterday, although I wasn't using dyndrivers in yesterday's testing)--a noteworthy system whose shared libs PLplot does not currently support--and obtained this: > ../configure No defaults file found, performing full configure. system is: SunOS-5.7 checking for prefix by location of plrender... not found -- using default checking for KCC... yes Found KAI C++, using that! checking for f77... no checking for g77... yes checking how to run the C preprocessor... cc -E checking for X... libraries , headers checking for main in -lX11... yes warning: gd header files not found checking for main in -lgd... no warning: gd library not found warning: cd header files not found checking for main in -lcd... no warning: cd library not found checking for tcl.h... /usr/local/include/tcl.h checking for libtcl... /usr/local/lib/libtcl8.2.so checking for itcl.h... no warning: can't find itcl.h, setting enable_itcl to no checking for tk.h... /usr/local/include/tk.h checking for libtk... /usr/local/lib/libtk8.2.so checking for matwrap... no warning: 'matwrap' not found, disabling Octave support. checking for Python.h... no warning: can't find Python.h, setting enable_python to no checking for main in -lXbsd... no checking for main in -lsocket... yes checking for main in -lnsl... yes checking for main in -lieee... no checking for main in -lm... yes checking how to make shared libraries... unknown checking how to make archive libraries... done checking for dynamic drivers... -n plmeta -n null -n tek -n dg300 -n ps -n xfig -n ljii -n hpgl -n ljiip -n impress -n pbm -n pstex checking for ANSI C header files... yes checking for unistd.h... yes checking for termios.h... yes checking for sys/wait.h that is POSIX.1 compatible... yes checking for ranlib... ranlib checking for pid_t... yes checking for vfork.h... no checking for working vfork... yes checking for popen... yes checking for usleep... yes checking for caddr_t... yes creating Makefile.in creating Makedemo.in creating drivers/drivers.db ../configure: test: argument expected > Whic is pretty bad. I think it is clear we have to leave dyndrivers off by default at this point in time. Alan, can you change it back? Joao Cardoso writes: > This implies that the shared libraries detection in sysloc.in should be moved > to the top of the file, and enable_dyndrivers should be set to no if shared > libs build is not supported in that system. This has to be done this > way, as other drivers configured afterwards depends on enable_dyndrivers. > > But this does not fully solve the problem, and as such I think that in the > RELEASE NOTES users should be warned that, is their systems shared libs are > not supported, they should configure without dyndrivers. |