ken@turtlevax.UUCP (Ken Turkowski) (11/19/85)
When I try compiling ttype.c with the -ltermcap flag, I get the following
message. What else do I need?
Undefined:
_initscr
_stdscr
_CA
__tty
__rawmode
__pfast
__tty_ch
__echoit
_LINES
_wmove
_wclrtoeol
_COLS
_subwin
_wclear
_wgetch
_wrefresh
_printw
_wstandend
_wstandout
_wclrtobot
_endwin
--
Ken Turkowski @ CIMLINC (formerly CADLINC), Menlo Park, CA
UUCP: {amd,decwrl,hplabs,seismo,spar}!turtlevax!ken
ARPA: turtlevax!ken@DECWRL.DEC.COMbrooks@leadsv.UUCP (Brooks Price) (11/20/85)
-lcurses -ltermlib should be included in the compile line... Brooks Price
jdg@husky.uucp (Jeff Gortatowsky) (11/22/85)
-=-=-=-==-=-=-=-=-=-= Low cal Line -=-=-=-=-=-==-=-=-=-=-=-=-=-=-=- > When I try compiling ttype.c with the -ltermcap flag, I get the following > message. What else do I need? > Undefined: > _initscr > Ken Turkowski @ CIMLINC (formerly CADLINC), Menlo Park, CA > UUCP: {amd,decwrl,hplabs,seismo,spar}!turtlevax!ken > ARPA: turtlevax!ken@DECWRL.DEC.COM Try -ltermcap AND -lcurses together... Jeff Gortatowsky {allegra,seismo}!rochester!ritcv!husky!jdg Eastman Kodak Company {eagle,astrovax,netword!}sun!sunrise!husky!jdg These comments are mine alone and not Eastman Kodak's or anyone else's at/in this relative space and time. -- Jeff Gortatowsky {allegra,seismo}!rochester!ritcv!husky!jdg Eastman Kodak Company {eagle,astrovax,netword!}sun!sunrise!husky!jdg These comments are mine alone and not Eastman Kodak's or anyone else's at/in this relative space and time.
garyjohn@hpspkla.UUCP (garyjohn) (12/02/85)
I also have a problem compiling ttype. Our system apparently does not have timeb.h. Could someone please post it? Thanks in advance. Gary Johnson
jpn@teddy.UUCP (12/03/85)
>I also have a problem compiling ttype. Our system apparently does not >have timeb.h. Could someone please post it? Thanks in advance. > Please DONT!!!!! If your system does not have timeb.h, then it also doesn't have ftime(), the function that uses the type "struct timeb". If you don't have this function, then you need a total rewrite of the program (if that is even possible)! ftime returns time in units SMALLER than one second. System III/V do not support this functionality (a major flaw, in my opinion) making timing tasks such as in ttype much more difficult. John P. Nelson, Moderator, mod.sources (decvax!genrad!panda!jpn seismo!harvard!wjh12!panda!jpn)
stephenf@elecvax.eecs.unsw.oz (Stephen Frede) (12/06/85)
>If your system does not have timeb.h, then it also doesn't have ftime(), >the function that uses the type "struct timeb". If you don't have this >function, then you need a total rewrite of the program (if that is even >possible)! ftime returns time in units SMALLER than one second. System III/V >do not support this functionality (a major flaw, in my opinion) making >timing tasks such as in ttype much more difficult. V7/SysV times(2) system call returns a value in tics (1/60 or 1/100 sec in the US, usually 1/50 sec in Australia) The ttype program does not use the value for anything very crucial. The following routine will probably be close enough. #ifdef BSD #include <sys/timeb.h> #else struct timeb /* only two elements used */ { time_t time; unsigned long millitm; }; #define index strchr #define rindex strrchr #endif /* BSD */ /* ... */ #ifndef BSD ftime(t) struct timeb *t; { struct tm tbuf; time(&t->time); t->millitm = (times(&tbuf) % HZ) * (1000/HZ); } #endif /* BSD */
garyjohn@hpspkla.UUCP (garyjohn) (12/06/85)
I received several responses to my (misguided) request for timeb.h, including explanations of why it wouldn't do me any good. It turned out that the patches needed to make the program run on our SysV system weren't very extensive, far from a "total rewrite". Of course, the calculated typing speeds are probably grossly inaccurate since the program now has only 1 second timing resolution, but I was more interested in a tool for practice than in precise evaluation of my skill, anyway. Many thanks to everyone who responded. Gary Johnson
thomas@utah-gr.UUCP (Spencer W. Thomas) (12/09/85)
In article <1929@elecvax.eecs.unsw.oz> stephenf@elecvax.eecs.unsw.oz (Stephen Frede) writes: >V7/SysV times(2) system call returns a value in tics (1/60 or 1/100 sec in >the US, usually 1/50 sec in Australia) >The ttype program does not use the value for anything very crucial. >The following routine will probably be close enough. This will not work because, while ftime returns REAL TIME in small units, times returns CPU TIME, obviously not the same thing, unless your process is using 100% of the CPU. -- =Spencer ({ihnp4,decvax}!utah-cs!thomas, thomas@utah-cs.ARPA) "Ask not what your kernel can do for you, but rather what you can do for you kernel!"
hansen@pegasus.UUCP (Tony L. Hansen) (12/09/85)
< If your system does not have timeb.h, then it also doesn't have ftime(), < the function that uses the type "struct timeb". If you don't have this < function, then you need a total rewrite of the program (if that is even < possible)! ftime returns time in units SMALLER than one second. System < III/V do not support this functionality (a major flaw, in my opinion) making < timing tasks such as in ttype much more difficult. To get < 1 second timing information on System III/V, one can make use of the times(2) system call. In addition to filling in a structure with information that is not needed for this purpose, it returns the elapsed real time, in 60ths or 100ths of a second depending on the processor, since an arbitrary (fixed per process) point in the past. By using this system call, I have been able to get the functionality needed for programs such as ttype. I'm not going to debate flawness issue. Tony Hansen pegasus!hansen
jsisi@trwrba.UUCP (Gerald W. Irmler) (12/10/85)
The referenced article stated the System V times call returned the CPU time. I haven't looked at the code, but I quote from the times(2) manual page: "Upon successful completion, *times* returns the elaspsed real time, in HZ, since an arbitrary time in the past (e.g. system start-up time). This point does not change from one invocation of *times* to another." Thus I think one can use times(2) to fake a suitable ftime unless the manual page is in error. Dan Jones ihnp4!trwrb!jcc-one!djones NOT the path in the header.