On Wed, Oct 29, 2008 at 11:44:31AM +0000, Daniel P. Berrange wrote: > This patch moves the definition for virRaiseError/virReportError out of > the internal.h file, and into a newly created virterror.h file, to match Hum, i was looking at this, yes using a new haeader makes sense, I will also move the libvirt error initialization to virterror.c But the problem is that we already have a virterror.h header which is the public one, I think adding a second one internally with the same name is a recipe for confusion. Let's name the internal one error.h (or virt_error.h) but not the same. > naming of the virterror.c file where the implementation lives. Most > code has to be updated to include virterror.h. This patch also drops the > leading __ prefix on these two functions - it was pointless, since we never > export it from the .so, not even for libvirtd use. > > The motivation for this change is to get clarity over our modularization. > Every file in our source tree ultimately includes internal.h, but not > everything is linking to the virterror.c file which includes the impl of > virRaiseError. So it is good to avoid defining functions in header files > which aren't available to those including the header. I think the name clash between the private and public header is more problematic than having a slightly different name for the .c and .h virterror.c implements the entry points for the public header virterror.h, like libvirt.c implements the entry points for the public header libvirt.h . having a second internel error header separate from internal.h is fine but we should just avoid the name clash. +1 for renaming to error.h or virt_error.h Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@xxxxxxxxxxxx | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/ -- Libvir-list mailing list Libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list