On Fri, May 10, 2013 at 12:30:39AM +0200, Jeff King wrote:
> On Thu, May 09, 2013 at 06:20:18PM +0200, Heiko Voigt wrote:
>
> > +static int config_buf_fgetc(struct config_source *conf)
> > +{
> > + if (conf->buf.pos < conf->buf.len && conf->buf.buf[conf->buf.pos])
> > + return conf->buf.buf[conf->buf.pos++];
> > +
> > + return EOF;
> > +}
>
> It probably would not matter for config reading, but is there any reason
> not to make your fgetc replacement 8-bit clean (i.e., to check only len,
> and not worry about NUL characters)?
I got a NULL character here and by mistake thought that somehow the len
would include the ending character. But that was a bug introduced by me
when I refactored to buffer and length instead of strbuf. I missed to
transport the len properly. So I will removed this check again.
Cheers Heiko
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html