On Tue, 2003-02-18 11:00:23 -0500, Andrew Rodland <arodland@noln.com>
wrote in message <b2tl9c$48c$1@main.gmane.org>:
> Jan-Benedict Glaw wrote:
> > On Tue, 2003-02-18 14:50:38 +0100, Tomas Szepe <szepe@pinerecords.com>
> > wrote in message <20030218135038.GA1048@louise.pinerecords.com>:
> >
> > This is the first time I really look at the code, so please forgive if I
> > talk about things where already a consens was given...
>
> >> +const unsigned char morsetable[] = {
> >> + 0122, 0, 0310, 0, 0, 0163, /* "#$%&' */
> >> + 055, 0155, 0, 0, 0163, 0141, 0152, 0051, /* ()*+,-./ */
> >> + 077, 076, 074, 070, 060, 040, 041, 043, 047, 057, /* 0-9 */
> >> + 0107, 0125, 0, 0061, 0, 0114, 0, /* :;<=>?@ */
> >> + 006, 021, 025, 011, 002, 024, 013, 020, 004, /* A-I */
> >> + 036, 015, 022, 007, 005, 017, 026, 033, 012, /* J-R */
> >> + 010, 003, 014, 030, 016, 031, 035, 023, /* S-Z */
> >> + 0, 0, 0, 0, 0154 /* [\]^_ */
> >> +};
>
> >
> > You're using a set bit for long and an unset bit for a short beep, don't
> > you? Storing these values in octal/as chars is quite low on memory
> > consumption, but I'd like to learn so I suggest:
>
> It's slightly more complicated than that:
> It's set bits for long, unset bits for short, and termination when the byte
> equals 0x01 (in other words, there's an extra set bit to the left of what
> we want). This lets us represent any variable-length morse of up to 7
> dits/dahs with a byte, which is cool because nothing is more than 6, that
> I've ever seen.
So you've got a leading I bit set. Morse alphabet has got 2, 3, 4, 5 or
6 dots/dashes, right? Then I vote for either having a macro like this:
#define MORSE(letter, len, b1, b2, b3, b4, b5, b6, b7) \
(1<<(len)| [all the shifting stuff here as of the last mail])
or let's have5 separate macros for all possible lengths (which are
possibly defined by using the above macro).
Really, I'm 100% for learning!
> The use of macros is an OK hack though, it reminds me of the nethack source.
> :)
>
> The reason someone proposed this in the first place is because I had had
>
> const unsigned char * morsetable [] = {
> ".-..-.", NULL, "...-..-"
>
> and so on in the initial revision of my patch, which is quite readable, but
> takes up a lot more space, and makes the code actually a bit messier too.
Oh, that's long, too...
What about
#define IS_DASH(letter, shift) \
((letter) == '-'? (1 << shift): (0 << shift))
MORSE(shift, b1, b2, b3, b4, b5, b6) \
(1 << (shift) | IS_DASH((b1), 5) | IS_DASH((b2), 4) \
| IS_DASH((b3), 3) | IS_DASH((b4), 2) \
| IS_DASH((b5), 1) | IS_DASH((b6), 0)
#define MORSE1(letter, b1) \
MORSE(1, '.', '.', '.', '.', '.', (b1)))
#define MORSE2(letter, b1, b2) \
MORSE(2, '.', '.', '.', '.', (b1), (b2)))
#define MORSE3(letter, b1, b2, b3) \
MORSE(3, '.', '.', '.', (b1), (b2), (b3))
#define MORSE4(letter, b1, b2, b3, b4) \
MORSE(4, '.', '.', (b1), (b1), (b3), (b4))
#define MORSE5(letter, b1, b2, b3, b4, b5) \
MORSE(5, '.', (b1), (b2), (b3), (b4), (b5))
#define MORSE6(letter, b1, b2, b3, b4, b5, b6) \
MORSE(6, (b1), (b2), (b3), (b4), (b5), (b6))
Then, you can have
const char morses[] = {
MORSE2('A', '.', '-'),
MORSE4('B', '-', '.', '.', '.'),
MORSE4('C', '-', '.', '-', '.'),
MORSE3('D', '-', '.', '.'),
MORSE1('E', '.'),
MORSE4('F', '.', '.', '-', '.')
...
};
That's going to take exactly the same memory in the compiled vmlinux
image, *and* it's really readable:-) Of course, gcc will optimize any
added "bloat" away...
MfG, JBG
--
Jan-Benedict Glaw jbglaw@lug-owl.de . +49-172-7608481
"Eine Freie Meinung in einem Freien Kopf | Gegen Zensur
fuer einen Freien Staat voll Freier Bürger" | im Internet!
Shell Script APT-Proxy: http://lug-owl.de/~jbglaw/software/ap2/
This archive was generated by hypermail 2b29 : Sun Feb 23 2003 - 22:00:22 EST