Re: New grub suitable for Etch
Hi Marc
I decided to use another aproach and saw that there was a mistake,
Len Sorensen sent to mailing list some suggestions and one of then was to
change this piece of code. Now we are reusing the buf reather than
memcpy`ing, in fact, buf didn`t even have any memory malloc`d.
This change was roughly tested and as far as I can see does not have any
side-effects.
On Monday 15 January 2007 18:13, Otavio Salvador wrote:
> Marc 'HE' Brockschmidt <he@ftwca.de> writes:
> > Hi,
> >
> > Otavio Salvador <otavio@debian.org> writes:
> >> @@ -139,17 +143,16 @@
> >> + {
> >> + int f_len = grub_strlen(buf) - grub_strlen(tmp);
> >> + char *def;
> >> -+ int a;
> >> -+ for(a = 0; a < f_len; a++)
> >> -+ grub_memcpy(&def[a], &buf[a], sizeof(char));
> >> ++ buf[f_len] = '\0';
> >> ++ def = buf;
> >> + safe_parse_maxint (&def, &entryno);
> >> + }
> >
> > I'm a bit confused by this part of the diff. I don't know what
> > grub_memcpy does exactly, but I guess it's roughly doing what memcpy
> > does. In that case, the new version is a significant change, are you
> > sure there are no side-effects?
>
> Leandro was the guy who ported this code to 0.97 and is the most
> experienced one on this part of code. I've added him on Cc and hope he
> comments on it.
>
> I wasn't the person who did this change. Leandro, can you comment on
> this change?
Reply to: