getdate.y grammar doubts

问题

http://www.freebsd.org/cgi/cvsweb.cgi/~checkout~/src/usr.bin/tar/Attic/getdate.y?rev=1.9.12.1;content-type=text%2Fplain;hideattic=0

I am trying to understand how yyTimezone is calculated in code below:

| bare_time  '+' tUNUMBER {
    /* "7:14+0700" */
    yyDSTmode = DSToff;
    yyTimezone = - ($3 % 100 + ($3 / 100) * 60);
}
| bare_time '-' tUNUMBER {
    /* "19:14:12-0530" */
    yyDSTmode = DSToff;
    yyTimezone = + ($3 % 100 + ($3 / 100) * 60);
}

How I understand is, lets say the timestamp is 2011-01-02T10:15:20-04:00; this means its 0400 hours behind UTC. So to convert it into UTC, you add 0400 hours to it and it becomes 2011-01-02T14:15:20. Is my understanding correct?

How is that achieved in the codeblock I pasted above?

回答1:

The input would encode the offset like -0400. The 0400 part of that would be returned as the tUNUMBER token (presumably holding an unsigned value). This token is matched by the grammar rules, and can be used as $3.

To get the actual offset in minutes from the value 400, you first have to split it up in two halves. The hours part can be obtained with $3 / 100 (ie. 4), and the minutes part with $3 % 100 (ie. 0). Since there are 60 minutes in an hour, you multiply the hours by 60, and add the minutes to that ($3 % 100 + ($3 / 100) * 60), which would give the value 240. Then all that's left, is to add the sign, and store it in yyTimezone.

After all that, yyTimezone will contain the timezone offset in minutes.

来源：https://stackoverflow.com/questions/7185737/getdate-y-grammar-doubts