george@vax2.ccs.cornell.edu (George R Boyce) (06/15/87)
Mike Clarkson commented about some bugs in the VMS flavor of GNU EMACS version 18.44. I am working with 18.46... re: quota limit not supported resulting in lost files I wrote the show-quota function and the quota.el routines that solves the problem for everything except write-region which did not call the same hook. I sent this to RMS who responded that it could be improved in certain ways that I never had time to implement. The "todo" list now has an item that says the vms version should write the file to a different filename and then rename it back once no errors have occured. I guess that would work too. The same bug exists under BSD unix systems running with quota limits turned on. re: check the version of the pty driver in use I don't see that the current distribution has any of the original pty work written by Prasad. But on the other hand the mailbox version of the routines don't work either, at least I can't get them to work. So what *is* the current status of subprocess control? PTY or MAILBOX?? Is anyone out there working on it? Did I just volunteer to clean it up? re: set-xon Yes set-xon is obsolete, but the set-input-mode function works. Other problems... I'm using VMS C V2.3 with VMS 4.4 (and the associated libraries) and had three problems. First the compiler bug checks on LREAD.C. The V2.2 compiler does not, so don't delete your old compiler. Second, stdio.h now defines _IOFBF which makes sysdep.c compile a call to setvbuf() which doesn't seem to exist. I simply changed the conditionals. Finally, screen IO seems to be slower. At least it *feels* slower as compared with 18.31 compiled with VMS C V2.2. But I haven't tracked the problem down yet. (I am posting this to comp.emacs and mailing it to vms-gnu-emacs) George Boyce, Academic Computing, Cornell University george@vax2.ccs.cornell.edu (128.84.252.100), george@crnlvax2.bitnet