mohamed@hscfvax.harvard.edu (Mohamed_el_Lozy) (03/12/88)
In a recent posting to comp.unix.ultrix, George Robbins writes: >The /usr/include/curses.h header file for Ultrix 1.2 contains a kludge >that is supposed to correct a problem with CRMOD action in curses. >It's not clear exactly what the situation is, but the "fix" keeps sc >from working correctly. I don't know whether the problem exists with >other releases of Ultrix. > >The fix is to include the following lines a the end of sc.h: > >#ifdef ultrix >#undef nl >#undef nonl >#define nl() (_tty.sg_flags |= CRMOD,_pfast = _rawmode,stty(_tty_ch, &_tty)) >#define nonl() (_tty.sg_flags &= ~CRMOD, _pfast = TRUE, stty(_tty_ch, &_tty)) >#endif > The problem is not confined to sc, but is general. The same patch had to be added to make bpatch (recently posted to comp.sources.unix) work under Ultrix 1.2. Probably the best fix is for the system administrator to add these lines to the end of the curses.h file if the problem exists on his system. The following shar contains a little program to test out whether the fix is needed, and whether it works. The correct output is: Line 1 Line 2 Line 3 while the bug manifests itself by producing: Line 1 Line 2 Line 3 I have found the bug, fixable as above, in both Ultrix 1.2 and 2.0. Feedback from those brave souls who have brought up 2.2 would be most welcome. -------------------Cut here------------------Cut here------------------- #! /bin/sh # This is a shell archive, meaning: # 1. Remove everything above the #! /bin/sh line. # 2. Save the resulting text in a file. # 3. Execute the file with /bin/sh (not csh) to create the files: # Makefile # bug.c # fix.h # This archive created: Sat Mar 12 09:19:47 1988 export PATH; PATH=/bin:$PATH echo shar: extracting "'Makefile'" '(150 characters)' if test -f 'Makefile' then echo shar: will not over-write existing file "'Makefile'" else sed 's/^ X//' << \SHAR_EOF > 'Makefile' Xall: original modified X Xoriginal: bug.o X cc -o original bug.c -lcurses -ltermcap X Xmodified: bug.o X cc -o modified -DMODIFIED bug.c -lcurses -ltermcap SHAR_EOF if test 150 -ne "`wc -c < 'Makefile'`" then echo shar: error transmitting "'Makefile'" '(should have been 150 characters)' fi fi # end of overwriting check echo shar: extracting "'bug.c'" '(251 characters)' if test -f 'bug.c' then echo shar: will not over-write existing file "'bug.c'" else sed 's/^ X//' << \SHAR_EOF > 'bug.c' X#include <curses.h> X X#ifdef MODIFIED X#include "fix.h" X#endif MODIFIED X Xmain() X{ X initscr(); X nonl(); X X mvaddstr(1, 0, "Line 1"); X mvaddstr(2, 0, "Line 2"); X mvaddstr(3, 0, "Line 3"); X refresh(); X X mvcur(0, COLS -1, LINES -1, 0); X endwin(); X exit(); X} SHAR_EOF if test 251 -ne "`wc -c < 'bug.c'`" then echo shar: error transmitting "'bug.c'" '(should have been 251 characters)' fi fi # end of overwriting check echo shar: extracting "'fix.h'" '(179 characters)' if test -f 'fix.h' then echo shar: will not over-write existing file "'fix.h'" else sed 's/^ X//' << \SHAR_EOF > 'fix.h' X#undef nl X#undef nonl X#define nl() (_tty.sg_flags |= CRMOD,_pfast = _rawmode,stty(_tty_ch, &_tty)) X#define nonl() (_tty.sg_flags &= ~CRMOD, _pfast = TRUE, stty(_tty_ch, &_tty)) SHAR_EOF if test 179 -ne "`wc -c < 'fix.h'`" then echo shar: error transmitting "'fix.h'" '(should have been 179 characters)' fi fi # end of overwriting check # End of shell archive exit 0
jpn@teddy.UUCP (John P. Nelson) (03/15/88)
>I have found the bug, fixable as above, in both Ultrix 1.2 and 2.0. Feedback >from those brave souls who have brought up 2.2 would be most welcome. Yup, the bug exists in Ultrix 2.2 as well. I have reported this bug to DEC support every time we get a new release. I assume that my bug reports are getting filed (in the circular file). It is clear that whoever made this "fix" at DEC had no understanding of curses whatsoever. As in the earlier posting, the bug can be fixed by replacing the definitions of the nl() and nonl() macros in curses.h with: #define nl() (_tty.sg_flags |= CRMOD,_pfast = _rawmode,stty(_tty_ch, &_tty)) #define nonl() (_tty.sg_flags &= ~CRMOD, _pfast = TRUE, stty(_tty_ch, &_tty)) - john nelson. P.S. I did reach someone at DEC who would listen (Thanks Chet!) My understanding is that they have no plans to fix this bug, as curses is now "owned" by some group in England, who is working on a port of the terminfo curses for Ultrix 2.4. I'm just pissed that my bug reports submitted for Ultrix 1.2 and 2.0 were ignored.