[haiku-commits] Re: haiku: hrev43576 - src/system/libroot/os
- From: Philippe Saint-Pierre <stpere@xxxxxxxxx>
- To: haiku-commits@xxxxxxxxxxxxx
- Date: Fri, 30 Dec 2011 00:14:51 -0500
Hi,
On Thu, Dec 29, 2011 at 11:13 PM, marcus overhagen <
marcusoverhagen@xxxxxxxxx> wrote:
> On Thu, Dec 29, 2011 at 8:16 PM, <stpere@xxxxxxxxx> wrote:
>
> > d2a423e: driver_settings: Replace type of bufferSize from size_t to
> ssize_t
> >
> > At various locations, checks are made so that bufferSize is >= 0, which
> never
> > occurs. CID 4197.
> > -put_string(char **_buffer, size_t *_bufferSize, char *string)
> > +put_string(char **_buffer, ssize_t *_bufferSize, char *string)
>
> I don't think this change is correct. whats the use in having a signed
> buffer size?
The _bufferSize seems to be misnamed. It should be "remaining buffer size"
and everytime something is written, it is substracted from this, and there
are various check to see if it became negative (which never occured).
Philippe
Other related posts:
