karl@osu-eddie.UUCP (Karl Kleinpaste) (02/06/86)
I have nearly finished with a port of GNU Emacs 17.45 to Very Vanilla USG UNIX System 5.0 on a VAX 780. Contrary to flammable mail I received when I first mentioned it to bug-gnu-emacs, the shortname hacks are not workable as distributed in 17.45, but it doesn't take too terribly much to make them fly. A number of other hacks were done due to the 5.0 cc objecting strenuously to double definitions of variables, plus an incompatibility or two having to do with nlist structures and the like. But on the whole, much closer to portable than I expected it to be, or than it appeared to be when I first started with this. I hope to send context diffs to RMS tomorrow evening. I have no idea how long it takes him to integrate changes of this type into his releases, so don't hold your breath. One item of concern to me is that I had to chuck the supplied crt0.c entirely; the initial pre-dump emacs would build OK, but dumped core immediately when run. I have used a slightly modified version of the system-standard crt0.s out of /usr/src/lib/libc/vax/csu, and that makes this particular piece of the change shaky as far as the publicly-available aspects of emacs goes. I intend to send RMS a plain diff (not a context diff) of that change, which should be sufficient and still avoid the legal hassles of needing the real crt0.s. I haven't even looked at the lisp code yet, so I don't know if there are BSDisms in there that need to be USGified. I hope there's not much, but I just don't know yet. -- Karl Kleinpaste