;;; -*- Mode:LISP; Package:USER; Base:10; Readtable:Common-Lisp; Patch-File:T -*- ;;; Patch directory for Tape version 26 ;;; Written 10-Nov-88 17:00:27 by keith (Keith Corbett) at site Gigamos Cambridge ;;; while running on Johannes Brahms from band 1 ;;; with Obsolete System 126.144, Obsolete ZWEI 126.29, Obsolete ZMail 74.17, Obsolete Local-File 76.0, Obsolete File-Server 25.0, Obsolete Lambda-Diag 18.1, Obsolete Unix-Interface 15.0, Obsolete Tape 26.4, Microcode 1762, SDU Boot Tape 3.14, SDU ROM 103, 10/17. (:OBSOLETE ((0 "Tape Loaded" "keith" NIL 2799308916) (1 "Tapemaster read/write operations had signals to non-existent error flavor." "keith" NIL 2799848134) (2 "Improve error notifications. CHECK-FOR-ERROR accepts :ERROR-TRACE argument, something to print if a tape error occurs and the trace variable TM:*NOTIFY-ON-ERRORS* is non-NIL. Also add this argument wherever CHECK-FOR-ERROR is called." "keith" NIL 2799849906) (3 "LMFL tape format file property lists were being written and read using whatever readtable happened to be bound at the moment. This could cause READ-FROM-STRING errors. Fix is: 1) In :READ-FILE-HEADER, to win at reading tape plists written in the past, try ZL and then CL readtables (if a read error occurs). If we still get a read error, say so before printing the error. 2) Always, from now on, write the plist in \"standard\" (ZL) format. This is best for downward-compatibility." "keith" NIL 2799852653) (4 " This should solve the problem that's been plaguing our backups. Sometimes the last file written on the tape is finished with one EOF, but the next file fails to fit, and the double EOF is (for unknown reasons) not being written to indicate logical EOT. On a new tape, the remainder of the tape appears to be blank, which you find out next time you try to verify or restore the tape. On COMPARE-FILES during a backup, this blows out the entire backup. I added error handling for that case, but not RESTORE-FILES, because the integrity of a restoration is not affected." "keith" NIL 2799872582) ))