henry@utzoo.UUCP (Henry Spencer) (09/19/84)
Tested and booted a new unix, now operational. The last of the DZs are now out of the system (although still physically resident in the cpu box), and all terminals are running on DHs. The paging mode of the tty driver has been improved to handle wraparound on long lines, to be temporarily turned off by a settable character, and to prompt with a settable character. (Suitable paging defaults will be integrated into setpage(1) in the next day or so; please don't hard-code your own meanwhile.) The inode-search code in the kernel has been modified to start where it left off, instead of at the beginning, which should eliminate the "pregnant pause" phenomenon in the editors and the accompanying transient heavy load. The kernel allocation code for blocks and inodes has been revised to maintain free counts, which however are not initialized or looked at by the relevant software yet (soon...). The tty driver has been modified so that backslash erase always puts a backspace into the document, regardless of the precise choice of erase character, if erblank mode is on (which it is by default for everybody). Nobody's perfect, there are two things wrong. The line-wrap code isn't right, it stops prematurely (it's harmless but definitely wrong). I know where the problem is, we'll fix this (probably) tomorrow. And, a bit more seriously, the 5121 line is out of order until tomorrow and will not be working as a dialin for the next few days. The DH driver doesn't have some of the DZ driver's kludgey dialout code yet, and until this is done 5121 will be dedicated to dialout and the alternate dialout (2581) won't dial out at all. Furthermore, until we get a jumper (software or hardware) on 5121, it won't dial out either, due to an oversight. Stay tuned. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry