Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!dali.cs.montana.edu!milton!mrc From: mrc@milton.u.washington.edu (Mark Crispin) Newsgroups: comp.sys.next Subject: Re: Nemacs and NeXT Message-ID: <1991Jun24.033405.25855@milton.u.washington.edu> Date: 24 Jun 91 03:34:05 GMT References: Organization: University of Washington, Seattle Lines: 50 In article mat@zeus.opt-sci.arizona.edu (Mat Watson) writes: >I need to make sure that I can port nemacs to the NeXT before >buying the machine. Has anyone used nemacs (wnn, and egg) on >their NeXTstation? DANGER DANGER Will Robinson! I don't use nemacs (not enough disk space), but I do use wnn as well as kterm under co-Xist on both a NeXTstation at home and a NeXTcube in the office. There is a bug in 2.0 and 2.1 that causes the NeXT kernel to *crash* when jserver (the server end of the Wnn4 system used by wnn, kinput, nemacs, etc.) is halted. This happens whether you kill jserver by hand, run wnnkill, or attempt to shut down or reboot the system. co-Xist has nothing to do with it; the problem is jserver which is completely independent of any X stuff. The result is that all the mounted filesystems are potentially marked as being "damaged" and fsck has to be run before they can be used again. It is annoying enough waiting for fsck to finish on a 105MB NeXTstation. It is infuriating to wait for it to finish on a 330MB + optical NeXTcube. I can only imagine what it'd be like for a n-GB user. I haven't lost any files yet, but I assume it's only a matter of time. I try to cover my ass by running sync a few times and swatting the machine immediately afterwards, with only partial success. I complained to NeXT about it in March when I first encountered it in 2.0 and they told me it's fixed in 2.1. It isn't; when I got my NeXTstation in April (with 2.1) I found it has the bug. I got an acknowledgement of my repeat bug report for 2.1 but that's all. If, as you suggest, nemacs is going to be a primary application on your new machine, you may want to consider getting a cheap SUN or DECstation instead of a NeXT. I can not recommend a NeXT for a Wnn4 user until and unless NeXT commits to fixing this bug and getting the fix out in the field. It is inexcusable that NeXT has been so careless with 2.1 (including shipping me a 68040 upgrade with the obsolete 2.0 two months after shipping a NeXTstation with 2.1). I'd give more details of the bug, but the blather scrolls past on the screen a mile a minute and is immediately erased when the machine automatically reboots. All I can see is that it is some kind of unexpected kernel page fault. It may be that this is to induce those of us who use Wnn and kterm freeware to cough up JPY60,000 (about $435) to Canon for 2.1J instead.