Julian Assange writes:
Ick. I presume you are refering to the heavy usage of the "%.128s" etc
construct. This notice that the field width is *after* the full-stop.
This limits the number of characters sprinted, but doesn't pad.
Can you perhaps try a test program under sunos-4.1.x to see what the
result is?
I'm not sure I've created the proper test. I tried
----------[ptest.c]----------
#include <stdio.h>
main()
{
char *longstring = "This is a very long string to see what happens to printf under Suns-4.1.3";
char *path = "news.connix.com/comp/infosystems/www/servers/unix/12345";
char buf[BUFSIZ];
int len;
len = strlen(longstring);
sprintf (buf, "%d %.255s\n%.348s", len, longstring, path);
puts(buf);
}
-----------[eof]-----------
which produces correct results
----------[ptest output]----------
73 This is a very long string to see what happens to printf under Suns-4.1.3
news.connix.com/comp/infosystems/www/servers/unix/12345
-----------[eof]-----------
Yet the entries for long pathnames in cache.history here are
truncated (the first is correct; the last two are lopped off):
-----------[cache.history]-----------
71 55v2iu$j87@lyra.csx.cam.ac.uk
news.connix.com/comp/text/tex/46852
82 55umc0$6uu@gold.interlog.com
news.connix.com/comp/infosystems/www/servers/un
83 55ts53$ni@gazette.bcm.tmc.edu
news.connix.com/soc/history/war/world-war-ii/21
-----------[eof]-----------
If you want to suggest a test that will more exactly duplicate the
code in nntpcache-1.0/history.c, I'll be glad to try it.
--
Ronald Florence Maple Lawn Farm, Stonington, CT
ron@mlfarm.com http://www.connix.com/~mlfarm
