deutsch@jplgodo.UUCP (Michael Deutsch ) (05/05/88)
These are questions for "C" and "UNIX" hackers. Your answer to all of these or a few will be greatly appreciated. 1. A UNIX function call equivalent to: "mkdir" command is??? 2. A programm running with "setuid" privilege, i.e.: -rwsr-xr-x tries to create directory and fails when executed by non-root, however, it works when executed by root, why??? 3. How to issue an "lp/lpr" command from the program without diagnostic popping up on the terminal??? 4. Which header file or type contains info on the file, i.e. (size, date created, date modified, number of links, etc.). 5. While using curses: one program spawns / forks/ issues a shell command to start program 2. When the program 2 finishes the control comes back to the parent process that tries to refresh the screen i.e. restore to the state prior to invoking program 2, what curses call will do this? I've tried refresh(curscr); but it does not seem to work????????????? 6. My program uses keypad, and that works fine with all UNIX - terminfo based systems; however pre 2.2 releases of XENIX were based on the termcap and did not have a: keypad( ) call. I can get around using keypad under XENIX; however why the program crashes if a keypad key is pressed under XENIX, how to get around this problems???? Please reply by e-mail, or call me collect at: (619) 452 - UNIX (8649) you may send a FAX as well: (619) 452 - 8711 Thanks. Michael Deutsch
allbery@ncoast.UUCP (Brandon S. Allbery) (05/15/88)
I saw some posted answers to these, many of which were at least somewhat inaccurate. As quoted from <14827@jplgodo.UUCP> by deutsch@jplgodo.UUCP (Michael Deutsch ): +--------------- | 1. A UNIX function call equivalent to: "mkdir" command is??? +--------------- BSD4.2/4.3 and System V Release 3 have mkdir(path, mode) system call. Under Xenix, you have to build a shell command to call /bin/mkdir and system() it. +--------------- | 2. A programm running with "setuid" privilege, i.e.: -rwsr-xr-x | tries to create directory and fails when executed by non-root, | however, it works when executed by root, why??? +--------------- Since /bin/mkdir is setuid, it runs as root/ORIGINAL USER. The setuid on mkdir overrides the setuid on the calling program. Thus, the ORIGINAL user must have permission to create the directory. The mkdir() system call doesn't have this problem. +--------------- | 3. How to issue an "lp/lpr" command from the program without | diagnostic popping up on the terminal??? +--------------- The "lp" command has an "-s" option to prevent the "request-id is ..." message from being displayed. If you actually mean error messages, you have to redirect standard error to /dev/null (lp ... 2> /dev/null). +--------------- | 4. Which header file or type contains info on the file, i.e. (size, | date created, date modified, number of links, etc.). +--------------- /usr/include/sys/stat.h. Look up stat(S). +--------------- | 5. While using curses: one program spawns / forks/ issues a shell command | to start program 2. When the program 2 finishes the control | comes back to the parent process that tries to refresh the | screen i.e. restore to the state prior to invoking program 2, | what curses call will do this? I've tried | refresh(curscr); | but it does not seem to work????????????? +--------------- refresh() doesn't take an argument and always works on stdscr. Use the call wrefresh(curscr) (note the initial "w"!). +--------------- | 6. My program uses keypad, and that works fine with all UNIX - terminfo | based systems; however pre 2.2 releases of XENIX were based on | the termcap and did not have a: | keypad( ) | call. I can get around using keypad under XENIX; however why | the program crashes if a keypad key is pressed under XENIX, how | to get around this problems???? +--------------- Crashes under which version of Xenix? -- curses/terminfo or termcap? There is a bug in some Xenixes with terminfo (Altos Xenix 5.2as1) where keypad keys are not properly recognized; programs don't crash, they simply get the escape sequence *with parity* instead of the KEY_* code. The curious thing about it is that when I was testing it, neither my port nor my terminal were configured for parity.... I'd have to see the program in order to figure out what was going on. -- Brandon S. Allbery, moderator of comp.sources.misc {well!hoptoad,uunet!marque,cbosgd,sun!mandrill}!ncoast!allbery Delphi: ALLBERY MCI Mail: BALLBERY