Gary> I keep getting a "(local-variables/warning) File Gary> local-variables error: Symbol's value as variable is void: Gary> compile-command" warning buffer appearing EVERYTIME I switch Gary> buffers. Is there a -*- compile-command: something -*- at the top of the Java file, or a /* Local variables: ... section containing compile-command at the bottom of the file? If not, that's quite a strange message. The only other thing I can suggest offhand is to confirm which java-mode you're using (it's probably from cc-mode) by finding a .java file, eliciting the bug, and typing C-h f java-mode RET to get its docstring. I hesitate to recommend bothering the CC Mode people yet if it is their version, but depending on how much of a hurry you're in, you could check with them. (See the docstring for how.) Gary> do it via XEmacs, got a message saying: "Cannot open load Gary> file: message" when I select the menu option "Mail -> Send Gary> and Exit". sigh...) message.el is part of Gnus, but we don't use that by default since it's a big package that many people see no reason to install. Is there any chance you, or somebody you borrowed customizations from, is using Gnus as their mail-user-agent? Gary> Microsoft Windows XP [Version 5.1.2600] *sigh* Gary> Building XEmacs in Gary> "E:\\acs\\software\\XEmacsWindowsKit\\xemacs-21.4.18\\nt". Using the prebuilt version? Windows has this habit of picking up the packager's defaults that get squirreled away somewhere in the preloaded files. That's _possible_ for `message' but doesn't seem likely to explain the Java thing. What does C-h v mail-user-agent RET have to say? If you restart XEmacs without loading the init file (you'll have to start it from a DOS window with --vanilla as far as I know), what does it say about mail-user-agent? Gary> Searching for program: No such file or directory, xargs Something's old, I think. We just fixed a bug with xargs (standard on Unix, not on Windows---we used to use it unconditionally, now we test for it ... maybe that's just a loud test?) -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.

vin> Please explain what you mean by 'picking up the packager's vin> defaults'. Can I avoid this by building without the HOME vin> envariable set? It sounds like the setup kit is contaminated vin> somehow and I want to understand how to avoid this. I actually doubt this is happening, but similar things have happened before. They're all arguably XEmacs bugs, but they're only going to bite in preinstalled systems. One case I can think of is the "abort, retry, reinstall Windows?" fiasco. What happened there was that Andy was building on a system with his sources on the F: drive, and something got built in. XEmacs would go searching for stuff at startup and try to access the F: drive. This is normally no problem, but Windows 2000 had this very aggressive preemptive caching problem, and if there was an empty removable media drive, it it didn't say "Oh, there's nothing there, let's fail and move on", it said "the program thinks something's there, we'd better get the user to put the media in". Another was some problems we had with the portable dumper. There were some settings for command line variables that got overwritten by the dumped values. This probably isn't a problem here, especially since I went through and audited all the command line variables to see which ones were used after the dumped code gets loaded. If this did regress, again it's especially likely to bite preinstalled system more often and harder. A third case would be a downstream packager problem. We've intermittently had problems with Debian because they try to ensure their standalone LISP packages work with both GNU Emacs and XEmacs, but sometimes they screw up and code that should only be executed for GNU Emacs gets executed for XEmacs. Bottom line is I have no suggestions for you, we need to find out what's going on here first. In any case if there is such a problem, it's an XEmacs bug IMO, not a problem in your procedure. There just shouldn't be any cruft in the binary. -- School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.