PERFORCE change 123376 for review

[Alexey Tarasov]

> http://perforce.freebsd.org/chv.cgi?CH=123376
>
> Change 123376 by taleks at taleks_th on 2007/07/12 14:27:18
>
> @@ -104,11 +107,11 @@
>       }
>
>       if ( (from == 0) && (size == 0) ) {
> -             sprintf(http_data, "GET /%s HTTP/1.1\r\nHost:
%s\r\nConnection: Close\r\n\r\n",
> +             sprintf(http_data, "GET /%s HTTP/1.1\r\nHost:
%s\r\nConnection: Close\r\nUser-Agent: pxe_http/0\r\n\r\n",
>                   filename, server_name
>               );
>       } else {
> -             sprintf(http_data, "GET /%s HTTP/1.1\r\nHost: %s\r\nRange:
%d-%d\r\nConnection: Close\r\n\r\n",
> +             sprintf(http_data, "GET /%s HTTP/1.1\r\nHost: %s\r\nRange:
bytes=%d-%d\r\nConnection: Close\r\nUser-Agent: pxe_http/0\r\n\r\n",
>                   filename, server_name, from, from + size
>               );
>       }

> You might want to change long lines like these to something as follows:
Ok, I'll make lines shorter.
I've not found strict rule in style(9) for ");" at the end of multiple lines
of parameters in function call, and thought my variant is more simple for
reading. Same intention is in using curly brackets in complex single
statement in if-else block. Well, it's not big problem, I may correct this
on next submits.

> And, don't use "sprintf". Use "snprintf" instead.
As for snprintf(), I'm also wanted to make it safer, but according to man
and header file - libstand have no such function. Same for strnstr().
I'm thinking about skipping sprintf() with many parameters while generating
header and writing directly to socket, in such case will be more function
calls and checks, but only  one of them (Range field, that always has
predictable size smaller then default buffer size) will use sprintf().

Best regards,
Alexey