开发者_StackOverflow中文版I'm seeing some weird behaviour when using wordexp() in a minimal C program when started within Xcode. I cannot reproduce this by starting the compiled binary from the command line.
#include <stdio.h>
#include <stdlib.h>
#include <wordexp.h>
#include <errno.h>
#include <assert.h>
int main (int argc, const char * argv[])
{
int wordexpResult;
wordexp_t words;
char* origPath = "~";
wordexpResult = wordexp(origPath, &words, 0);
printf("wordexpResult = %i\n", wordexpResult);
switch (wordexpResult)
{
case 0:
break;
case WRDE_BADCHAR:
fprintf(stderr, "BADCHAR\n");
exit(EXIT_FAILURE);
case WRDE_BADVAL:
fprintf(stderr, "BADVAL\n");
exit(EXIT_FAILURE);
case WRDE_CMDSUB:
fprintf(stderr, "CMDSUB\n");
exit(EXIT_FAILURE);
case WRDE_NOSPACE:
fprintf(stderr, "NOSPACE\n");
exit(EXIT_FAILURE);
case WRDE_SYNTAX:
fprintf(stderr, "SYNTAX\n");
exit(EXIT_FAILURE);
default:
fprintf(stderr, "Unrecognized value: %d\n", wordexpResult);
exit(EXIT_FAILURE);
}
assert(words.we_wordc != 0);
wordfree(&words);
return 0;
}
Sometimes (about one in five runs), the assertion evaluates to false, even though wordexp() always returns 0 (i.e. none of the cases in switch() are executed). This means that wordexp() does not return an error, but it does not expand the tilde either. How can that be?
Luckily, I don't have to rely on wordexp right now, but I'd still be interested in what is going on here. I mean... there's no multi threading, no varying input data, nothing. Any ideas?
This is running on Mac OS X 10.6.4, Xcode 3.2.4.
Cheers,
Marco
精彩评论