svn commit: trunk/busybox/util-linux/dmesg.c "attribution"

[Rob Landley]

On Friday 28 July 2006 3:55 am, Mike Frysinger wrote:
> On Thursday 27 July 2006 12:40, landley at busybox.net wrote:
> > get the attribution right this time.
> 
> the copyright line was not for attribution, it was copyright
> 
> and it was correct in my commit as the code i added i wrote from scratch and 
> ive been assigning my work to Erik in busybox/uclibc/buildroot

You signed a piece of paper handing over your copyrights and Erik has it on 
file?  Because legally, you need a written instrument of conveyance (which is 
why the FSF has been such sticklers about that all these years, and is one of 
the central issues in the SCO vs Novell litigation groklaw's been covering 
for the past few years).  If you don't have that then mis-attributing a 
copyright can screw up our attempts to enforce the copyrights, opening us to 
procedural challenges and forcing us to do extra work to prove standing and 
such.  (Not that we can't do this, but it's definitely not helpful.)

As for the copyright on _this_ code, I checked the submission log and saw the 
code Bernhard Fischer added in 14688.  According to svn annotate:

>  14688      aldot       c = '\n';
>    357       erik       for (i = 0; i < n; i++) {
>  14688      aldot               if (c == '\n' && buf[i] == '<') {
>    357       erik                       i++;
>    357       erik                       while (buf[i] >= '0' && buf[i] <=
> '9') 357       erik                               i++;
>    357       erik                       if (buf[i] == '>')
>    357       erik                               i++;
>    357       erik               }
>  14688      aldot               c = buf[i];
>  14688      aldot               putchar(c);
>    357       erik       }
>  14688      aldot       if (c != '\n')
>    357       erik               putchar('\n');
>   6714   andersen all_done:
>  14688      aldot       if (ENABLE_FEATURE_CLEAN_UP) {
>   6714   andersen               free(buf);
>   6714   andersen       }

That came first, and that's what I was looking at when I was trying to get the 
size down.  (I tried several ways to modify the buffer (and length) in place 
so I could fall through to the one big write() call (which should really be a 
safe_write() but I forgot to fix that in the final checkin).  But the closest 
I got size-wise was 10 bytes larger, so I stuck with an implementation based 
on Bernhard's version because dmesg really isn't speed-critcial.)

Rob
-- 
Never bet against the cheap plastic solution.