[MPlayer-dev-eng] [PATCH 1/4] String handling audit/cleanup take 2

[Nicholas Kain]

> The +1 is a bugfix and will be applied. About the rest it's the same as
> in realrtsp: buffer was just created with the right size, why checkit
> again while writing into it?

It was just changed to make it a little easier on people who might
check the code for correctness later.  If the functionally equivalent
changes are unwanted, they can just be dropped.  I personally
like them because it makes code review a lot easier, but it's
not a big issue.

> >        npkt++;
> >      } else
> >        line = strdup(src_line);
> > @@ -526,12 +526,9 @@ parse_smil(play_tree_parser_t* p) {
> >                mp_msg(MSGT_PLAYTREE,MSGL_V,"Error parsing this source
> > line %s\n",line); continue;
> >              }
> > -          if (s_end-s_start> 511) {
> > -            mp_msg(MSGT_PLAYTREE,MSGL_V,"Cannot store such a large
> > source %s\n",line);
> > -            continue;
> > -          }
> > -          strncpy(source,s_start,s_end-s_start);
> > -          source[(s_end-s_start)]='\0'; // Null terminate
> > +          size = s_end-s_start+1;
> > +          source = malloc(size);
> > +          strlcpy(source,s_start,size);
>
> Probably ok, but that's an unrelated functional change (remove the
> 512 bytes line limit). this may be desiderable since i have a sample
> with a smil file on a single line (unsupported for other reasons atm).

Yeah, if there's an instance there this change causes breakage,
it can just be dropped.  I only converted it to use malloc() because
it seemed to be an easy fix/improvement to make while I was
there.