guest@albert.ai.mit.edu (Guest Account) (05/26/90)
In article <TOMW.90May23002947@orac.esd.sgi.com> tomw@esd.sgi.com writes: >In article <8577@life.ai.mit.edu> guest@albert.ai.mit.edu (Guest Account) writes: >> Here is the best shell I have ever used on MINIX. > >I just talked to Brian Fox, and he would really appreciate it if you >would send him diffs for whatever you changed. He also wants to know >what version of bash you ported because if it's for an old version, he >would like you to try it with a newer one. > I did a very quick job of the port (one weekend). There are no diffs, I didn't keep track of what I did. There weren't many changes, just a gcc-crash-elvis-gcc cycle. Whenever there was a missing library routine I just made up a fake one (ulimit is Very Fake <I hate ulimit anyway>) The only real tough part was simulating fcntl(..., NDELAY). I put in a routine which starts up a child, counts to 20 or so and kills it. The child reads one char and calls exit(ch). The parent examines the exit status. I call the function cr*ppy_read(), fortunately it does not get called in real life very much (at all?). The sources were 1.05 + patches off of prep.ai.mit.edu. I may have time to upload the whole thing, but I use kermit over telnet through many gateways to get here, so it's painfully slow.