racine@yunexus.yorku.ca (Jeff Racine) (04/02/91)
I wish to be able to set an environment variable (the path to a file)
and then to execute a program using the system() function in the
following fashion:
char *s;
s=getenv("FOO");
system("FOO/FUBAR");
or something like that. I am using Turbo-C running under MS-DOS, and
the manuals do not seem to help me :-).
I know this is a simple problem and that I am missing something
obvious (such as I should not be using the system() function). I would
really like some help on this approach (or rather the correct
aproach).
Any help would be greatly appreciated.
--------------------------------------------------------------------
Jeff Racine
racine@nexus.yorku.ca
racine@yunexus.UUCPjinx@milton.u.washington.edu (Everyone Knows JINX) (04/02/91)
racine@yunexus.yorku.ca (Jeff Racine) writes: >I wish to be able to set an environment variable (the path to a file) >and then to execute a program using the system() function in the >following fashion: > char *s; > s=getenv("FOO"); > system("FOO/FUBAR"); >or something like that. I am using Turbo-C running under MS-DOS, and >the manuals do not seem to help me :-). >I know this is a simple problem and that I am missing something >obvious (such as I should not be using the system() function). I would >really like some help on this approach (or rather the correct >aproach). To do this, I always did something like: s = getenv("FOO"); sprintf (command_line, "%s/FUBAR", s); system(command_line); It uses an extra variable, but...
jwm712@unhd.unh.edu (Jonathan W Miner) (04/02/91)
In article <22200@yunexus.YorkU.CA> racine@yunexus.yorku.ca (Jeff Racine) writes: >I wish to be able to set an environment variable (the path to a file) >and then to execute a program using the system() function in the >following fashion: > > char *s; > s=getenv("FOO"); > system("FOO/FUBAR"); > > [ stuff deleted ] >Jeff Racine >racine@nexus.yorku.ca >racine@yunexus.UUCP The following may do what you want: char *s; char temp[256]; s=getenv("FOO"); sprintf(temp,"%s/FUBAR",s); system(temp); Better ways probably exist, take a look at the exec functioins. Jon. -- Jonathan Miner | I don't speak for UNH, and UNH does not jwm712@unhd.unh.edu | speak for me! (603)868-3416 | Hacking to survive......
brendan@cs.widener.edu (Brendan Kehoe) (04/02/91)
In <22200@yunexus.YorkU.CA>, racine@yunexus.yorku.ca writes: >I wish to be able to set an environment variable (the path to a file) >and then to execute a program using the system() function in the >following fashion: > > char *s; > s=getenv("FOO"); > system("FOO/FUBAR"); Try: #include <stdio.h> #include <stdlib.h> #include <string.h> #define PROGRAM "FUBAR" main(argc, argv) int argc; char **argv; { char *s, *t; if ((s = getenv("FOO")) == (char *)NULL) { fprintf(stderr, "%s: no such variable FOO\n", *argv); exit(1); } t = (char *) malloc(strlen(s) + strlen(PROGRAM) + 1); sprintf(t, "%s/%s", s, PROGRAM); if (system(t) < 0) { perror("system"); exit(1); } } No guarantees -- I just typed it, didn't try it. :-) Brendan -- Brendan Kehoe - Widener Sun Network Manager - brendan@cs.widener.edu Widener University in Chester, PA A Bloody Sun-Dec War Zone Now that we know he has ID, we could give him an account. finger bush@cs....
cschmidt@lynx.northeastern.edu (04/07/91)
> I wish to be able to set an environment variable (the path to a file) > and then to execute a program using the system() function in the > following fashion [under DOS]: > > char *s; > s=getenv("FOO"); > system("FOO/FUBAR"); This brings to mind the following terrific idea. The DOS environment can be used to implement "logical names" in a manner similar to logical names on VMS. I have been doing this for years with great success. All my DOS programs allow you to use "logical names" in place of actual directory names and file names. To illustrate how this technique works, suppose I were to define a symbol as follows: SET H=C:\TC\INCLUDE\*.H or SET I=C:\TC\INCLUDE\ or SET TC=C:\TC\ I could then use my DOS File Manager program to access header files in the directory C:\TC\INCLUDE\ by typing: FM H: or FM I:.H or FM TC:\INCLUDE\.H The rule is that wherever a colon appears in a file specification, if the token to the left of the colon is an environment symbol, then the colon and the symbol are replaced by the translation. It's basically just a macro processor. The colon can be followed by the remaining portion of a complete path or file specification. Alternatively, the logical name itself might define a complete file specification. Syntactically, the logical name is identical to the drive name portion of a complete DOS file specification, except that logical names are not limited to a single letter. If the logical name is found in the DOS environment, the logical name and the colon are replaced by the translation; otherwise, the logical name and the colon remain in place, so actual drive names are not affected. Logical names can be defined in terms of other logical names. Logical name translation is repeated until no candidate logical name is found in the resulting translation (signaled by the absence of a colon), or until no translation is found in the DOS environment, or until the maximum number of iterations is reached. A maximum of eight iterations avoids looping indefinitely in the case of circular definitions. Wherever translation produces consecutive backslashes, one of the backslashes is deleted, which is helpful when referring to nested directories. Returning to the original problem, if you write a function named "translate" or whatever that returns a pointer to a string containing its input parameter with logical names translated, the solution to your problem might look something like this: system (translate ("FOO:FUBAR")); (In practice, I do not usually approve of functions that return pointers to internal static strings, but that is another issue. Also, an application with a hard-coded logical name should probably test whether that logical name is defined before trying to use it.) As an additional advantage, this technique is portable to VMS and Unix (I think). I do know that Unix has environment variables, and I know of no conflict that might result from using the colon this way in Unix file specifications. (Corrections on this point will be gratefully received.) Logical names seem much simpler to me that Unix "file links", which are usually used to accomplish the same thing. (Logical names are more powerful in some respects than links, and vice versa.) I think Microsoft should provide full support for logical names as defined here, in both DOS and OS/2 and in their compiler runtime libraries, and of course in their application products. Microsoft could then phase out the numerous, cumbersome, less powerful features they added to DOS to accomplish the same thing. Programmers and users have always been dissatisfied with the methods DOS provides for programs to find their data files. The sheer number of methods that have appeared in successive versions of DOS indicates the failure of any one method to provide an adequate solution. The logical name method is both simpler and more powerful than all other methods combined. If all DOS programs supported logical names, the DOS commands APPEND, ASSIGN, JOIN, and SUBST would never be needed. Users would be provided with a single, versatile method that would work across application, and across platform as well. In the meantime, you and your customers will benefit if you add logical name support to all your DOS programs. I would like to know what you think about that. It has worked wonderfully well in the many DOS applications I have written. Christopher Schmidt cschmidt@lynx.northeastern.edu
ckp@grebyn.com (Checkpoint Technologies) (04/08/91)
In article <memo.895702@lynx.northeastern.edu> cschmidt@lynx.northeastern.edu writes: >> I wish to be able to set an environment variable (the path to a file) >> and then to execute a program using the system() function in the >> following fashion [under DOS]: >> >> char *s; >> s=getenv("FOO"); >> system("FOO/FUBAR"); > >This brings to mind the following terrific idea. The DOS environment >can be used to implement "logical names" in a manner similar to >logical names on VMS. I have been doing this for years with great >success. > [ describes a way to assign a path spec to a drive letter. ] Yes yes yes. I work with many machines with assignments (most every DEC operating system, AmigaDOS) and they are great. I also work with Unix (though not DOS) and have found ways to survive using environment variables instead. I prefer to have assigns than not have them. -- First comes the logo: C H E C K P O I N T T E C H N O L O G I E S / / ckp@grebyn.com \\ / / Then, the disclaimer: All expressed opinions are, indeed, opinions. \ / o Now for the witty part: I'm pink, therefore, I'm spam! \/