conspy applet

[Denys Vlasenko]

On Mon, Jun 21, 2010 at 10:56 AM, Pascal Bellard
<pascal.bellard at ads-lu.com> wrote:
> Hello,
>
>> I added conspy to miscutils (util-linux does not seem to be the right
>> place).
>
> A 159 bytes (i386) shrink for conspy applet.

Applied, thanks.

I noticed a problem by "fb+512char font" combination.
I use this program to print all colors:

#!/bin/sh
printf "\033[0m" # clear all
for i in 0 1 2 3 4 5 6 7; do
    for j in 0 1 2 3 4 5 6 7; do
        printf "\033[4${i};3${j}m<>"
    done
    printf "\033[1m" # bright text
    for j in 0 1 2 3 4 5 6 7; do
        printf "\033[4${i};3${j}m<>"
    done
    printf "\033[0m\n" # clear all, and go to next line
done

and it looks ok on the console, but conspy's
colors are all messed up. It does not seem to be a conspy
fault. Rather, the attribute bytes in /dev/vcsa are wrong:

# cat /dev/vcsa2 | hexdump -C

and I see:

00000000  41 c8 20 08 3c 00 3e 00  3c 08 3e 08 3c 04 3e 04  |A. .<.>.<.>.<.>.|
00000010  3c 0c 3e 0c 3c 02 3e 02  3c 0a 3e 0a 3c 06 3e 06  |<.>.<.>.<.>.<.>.|
00000020  3c 0e 3e 0e 3c 10 3e 10  3c 18 3e 18 3c 14 3e 14  |<.>.<.>.<.>.<.>.|
00000030  3c 1c 3e 1c 3c 12 3e 12  3c 1a 3e 1a 3c 16 3e 16  |<.>.<.>.<.>.<.>.|
00000040  3c 1e 3e 1e 20 0e 20 0e  20 0e 20 0e 20 0e 20 0e  |<.>. . . . . . .|
00000050  20 0e 20 0e 20 0e 20 0e  20 0e 20 0e 20 0e 20 0e  | . . . . . . . .|

Let's look at second "<>" one which is red text on black.
In /dev/vcsa2 dump, it's 3c 08 3e 08.
But "red test on black" should be 01, not 08!
Then green one has 04 instead of 02.
And so on.

Looks like attribute bytes have their bit order reversed!

On my home machine where I have real text console (not framebuffer),
conspy colors look correctly. I also have 256-char font there.

I don't know whether it is caused by fb or by 512-char font.
I suspect fb. Will experiment a bit more.
-- 
vda