MKSTR(1) — UNIX Programmer’s Manual
NAME
mkstr − create an error message file by massaging C source
SYNOPSIS
mkstr [ − ] messagefile prefix file ...
DESCRIPTION
Mkstr is used to create files of error messages; it reduces the number of error diagnostics and therefore, reduces system overhead in running the program, as the error messages do not have to be constantly processed in and out.
Mkstr processes each of the specified files, placing a massaged version of the input file in a file with a name which consists of the specified prefix and the original name. A typical usage of mkstr would be as follows:
mkstr pistrings xx \(**.c
This command causes all the error messages from the C source files in the current directory to be placed in the file pistrings and processed copies of the source for these files to be placed in files with names prefixed with xx.
To process the error messages in the source to the message file, mkstr keys on the string ‘error("’ in the input stream. Each time it occurs, the C string starting at the ‘"’ is placed in the message file, followed by a null character and a new-line character. The null character terminates the message so it can be easily used when retrieved; the new-line character makes it possible to sensibly cat the error message file to examine its contents. The massaged copy of the input file then contains a lseek pointer into the file which can be used to retrieve the message, for example:
charefilname[] = "/usr/lib/pi_strings";
intefil = -1;
error(a1, a2, a3, a4)
{
char buf[256];
if (efil < 0) {
efil = open(efilname, 0);
if (efil < 0) {
oops:
perror(efilname);
exit(1);
}
}
if (lseek(efil, (long) a1, 0) || read(efil, buf, 256) <= 0)
goto oops;
printf(buf, a2, a3, a4);
}
The optional − causes the error messages to be placed at the end of the specified message file for recompiling part of a large mkstred program.
SEE ALSO
AUTHORS
William Joy and Charles Haley
3rd Berkeley Distribution — %W%%Q%%Y%