Manuel Menal, le Sun 31 Oct 2010 23:13:18 +0100, a écrit :
> I've included a change for the documentation as well.
Could you avoid reformating paragraphs? It makes finding out the actual
changes tedious.
Samuel
On 14/10/2010 19:10, Roland McGrath wrote:
>> Shouldn't we change io_write() too?
> Yes, they are exact parallels.
Here is a patch that should make io_read() and io_write() return ESPIPE
when called with offset != -1 for non-seekable objects. I've tested all
the components affected by this pat
> Shouldn't we change io_write() too?
Yes, they are exact parallels.
On 09/10/2010 21:00, Roland McGrath wrote:
> I think it would be reasonable to change the specification for io_read to
> say that the offset parameter must be exactly (off_t)-1 for a nonseekable
> io object, and that otherwise it must fail with ESPIPE. We might need to
> audit the existing uses to
> Manuel Menal, le Sat 09 Oct 2010 20:40:26 +0200, a écrit :
> > There should be a way to check if the object is seekable or not before
> > calling io_read(), so pread() can return an error. But there is no
> > io_seekable() RPC, and calling io_seek() seems wrong to me, since
> > pread() is suppose
I think it would be reasonable to change the specification for io_read to
say that the offset parameter must be exactly (off_t)-1 for a nonseekable
io object, and that otherwise it must fail with ESPIPE. We might need to
audit the existing uses to make sure there aren't uses that would violate
thi
Manuel Menal, le Sat 09 Oct 2010 20:40:26 +0200, a écrit :
> There should be a way to check if the object is seekable or not before
> calling io_read(), so pread() can return an error. But there is no
> io_seekable() RPC, and calling io_seek() seems wrong to me, since
> pread() is supposed not to c
