The 'gv' package appears to have been dropped from the FC4 tree. This is not okay, because no functional alternative has been incorporated. ggv, which appears to have morphed into evince, lacks an absolutely critical feature for document authors: the ability to "watch" a postscript file and reload when the postscript file has been modified.
I pulled the SRPMS from FC3 to see if it would rebuild cleanly and run correctly. The only edit required was to change the Copyright: tag in the SPEC file to a License: tag. The legacy Copyright: tag is no longer supported by rpmbuild.
It doesn't sound "critical" for me... Would you please file an upstream request (bugzilla.gnome.org) on evince about this?
You might notice that I didn't file a high-priority bug report... But let me explain why I think it is important. I think that you and I are making different use-case assumptions. evince is a much better tool for the general end user who is not engaged in document production. It is essentially "option free" and therefore relatively idiot proof. In my opinion, it makes absolutely crappy use of screen real estate, but most end users won't care -- they probably feel that visual consistency with other viewers is more important, which is fine. evince is not effective for document authors, because it doesn't fit into the document author's equivalent to the "edit compile debug" loop. It lacks the control options on things like bounding box handling, antialiasing, and change watching that gv had, but the most critical thing is --watch, (which is also the reason I never used ggv). Why? As an author, I sit with two things on my screen: gv on one side set to watch, and emacs on the other. I edit my latex document, run 'make' to generate a new PS file, and gv automatically reloads as needed. This lets me get near-wysiwyg in an authoring environment that can still be used during a build process. Also, in the only authoring environment that provides really high-quality formatting. I can't use things like lyx, because they don't provide sufficiently fine-grain macro support. I can't use things like OpenOffice because (a) the range of output types is insufficient and the template model is wrong for production purposes, and (b) there is no command line formatting capability that does not involve an interactive UI. So: the feature set of 'gv' probably isn't relevant to the end user, but it is nearly as important to a high-quality document author as "make" probably is to you. For this reason, I really want to see it back in the release. When evince provides equivalent control, by all means drop gv. But I have to say that if evince is targeted at general end users, adding this kind of feature load is precisely the wrong thing to do to evince...
I see yor point now and I agree with it. tetex is kept in Fedora Core because documentation writers (including Red Hat's) use it. For the same reason and based on above comment, I also sustain adding gv back to Core . However, I still think the --watch feature should be requested in bugzilla.gnome.org. Please list there the link to this bug and let the evince author decide if it's too much feature load. As it's a command line option, most users won't notice it.
gv isn't going to be added back to Core. If you're interested in maintaining it in Fedora Extras, please see http://fedoraproject.org/wiki/Extras for information on adding and maintaining packages in Fedora Extras