jspear@BLACKBIRD.AFIT.AF.MIL (Jon L. Spear) (11/12/88)
The GNU Emacs 18.52 distribution includes support for the ELXSI computers under Unix System V. ELXSI is mentioned in etc/MACHINES as: >Elxsi 6400 (m-elxsi; s-usg5-2.h) > > 18.36 required one small change that's installed in 18.38. > > m-elxsi.h currenty specifies CANNOT_DUMP. > The Elxsi can in principle dump, but the necessary changes to > unexec, which involve byte-swapping, were too ugly to install. > If someone submits simple code that supports dumping on the Elxsi, > it can be installed and CANNOT_DUMP can be turned off. I'll presume that this works and is useful to a segment of ELXSI users, but for those sites such as ours that choose to run BSD 4.3, it doesn't work (emacs doesn't even link). Trying really simple things like switching the system file to BSD resulted in linked executables that would either core dump or immediately grind to a dynamic halt. I tried to get some help from our Unix wizards in fixing it, but they were initially mystified and later couldn't get excited about trying to get their arms around something as large as GNU Emacs. In an appeal to comp.emacs, however, I got some very useful help from Matt Crawford (matt@oddjob.uchicago.edu), who sent me his patches to the 18.52 distribution. They appear to work just fine, and include the ability to produce a DUMPed emacs with pre-loaded lisp functions. Matt's patches seem like they would be appropriate for inclusion in the general GNU distribution, but... Matt said he has sent his patches to the GNU project for inclusion in future releases, but they appeared to have fallen on deaf ears (eyes?) as he got no reply. My questions (finally) are: 1) Are you willing to provide improved support for ELXSI computer users, specifically those running BSD 4.3? (assuming Yes,) 2) Do you already have Matt's patches to crt0.c, unexec.c, and config.h (should probably be integrated into m-elxsi.h)? 3) If so, are they in a suitable form for integration with the standard distribution, or should they be resubmitted in some other form? 4) If not, in what form shall we send them (diff files, complete source modules)? Thanks for all your work on GNU Emacs (the only GNU product I currently use). It's a great program. Keep up the good work! One other request though. Please reconsider your stance on non-support for Apple products. I fully agree that this country has far too many lawyers, that Apple has too many in particular, and Apple seems to be using its lawyers in questionable ways. This plague of lawyers and litigation may very well be responsible for not only the decline of the American Empire, but also for stifling innovation and improvements in the computer industry. We'll have to wait and see. In the mean time, many people (myself included) have bought Apple computers and find them to be useful tools at an acceptable price. Sure, a small part of that price goes to support questionable legal actions, but from what I've heard it seems possible that Apple's motivation and goals are not all that sinister. In any event, neither their legal maneuvering (for now) nor your declaration of non-support are likely to have much effect on sales of Apple computers. GNU non-support seems to me like more of an inconvenience to Apple computer users than to the Apple computer corporation. So, please reconsider a policy that appears to miss its real target. Just my opinion, of course. -Jon -- ---- random rumors from Jon Spear, comp eng student | With computers we can make a jspear@blackbird.afit.af.mil or @afit-ab.arpa | billion mistakes every second.