jack@cadre.dsl.PITTSBURGH.EDU (Jack Nelson) (03/06/87)
Has anyone gotten the mod.sources version of jove to use the interactive-shell feature? The version distributed with 4.3 seems to work in pipe mode on 2bsd; it won't work in ptys mode as far as I can take it. The mod.sources version seems to lack the feature. Is the 4.3 version more recent? Does anyone have a path to J. Payne? I would like to keep current with any patches that come out, but as it is the 4.3 version is better. UUCP: { akgua | allegra | cmcl2 | idis | ihnp4 | mi-cec | pitt | psuvax1 | sun | sunrise | vax135 } ! cadre ! nelslab ! jack ARPA: jack@cadre.PITTSBURGH.EDU John P. Nelson, M.D., 3811 O'Hara St, Pittsburgh, PA 15213, t:412-624-1769 Dept. of Psychiatry, U. of Pittsburgh
jpayne@rochester.ARPA (Jonathan Payne) (03/06/87)
The version of JOVE that was sent over mod.sources was supposedly the latest and greatest. If JOVE did pty's in the version that comes with 4.3, it should work fine in the later versions. What doesn't work about it? Since I sent JOVE to mod.sources I've gotten several bug fixes (10 times over ...) and a few enhancements. I've fixed everything and incorporated most of the enhancements. Often I've tried to reply to the people sending the fixes, but the mail bounced. So if you're wondering why I never reply that's probably why... I'm also wondering how to make interactive processes work better on versions of UNIX that don't have the SELECT system call. What's THE right way to do things? The way I did it once was to have EVERYTHING come in over a single pipe, including input from the keyboard. There were initially two processes, JOVE and the kbd process. The kbd process read characters and passed them on to JOVE with a little header identifying it. Subshells just used the same fildes, so JOVE was sitting there reading one fd ... it just worked. Main problem was that it was too slow for keyboard input. v7/2.9BSD pipes were implemented with disk files. It would probably work well with systems with sockets, since they use incore buffers. But those systems also have SELECT, which is the best way, anyway. So what's the right thing to do? Anyway, this is my path. jpayne@rochester.arpa, or jpayne@cs.rochester.edu or something like that. I think rochester talks to seismo via uucp ...