On Fri, 2008-09-26 at 17:23 +1200, Amos Jeffries wrote:
> >> Which makes more sense then to cleanup make flushReads() == kickReads(-1),
> >> or similar instead of duplication.
> >
> > Avoiding duplication was not "possible" when the class was written
> > because of the reentrant callbacks and lac
Alex Rousskov wrote:
On Fri, 2008-09-26 at 13:35 +1200, Amos Jeffries wrote:
On Thu, 2008-09-25 at 13:43 -0600, Alex Rousskov wrote:
I can probably fix it myself, but it would help a lot if somebody
could
document (briefly!) the overall purpose of deferred reads and the
exact
purposes of these
On Fri, 2008-09-26 at 13:35 +1200, Amos Jeffries wrote:
> > On Thu, 2008-09-25 at 13:43 -0600, Alex Rousskov wrote:
> >>
> >> I can probably fix it myself, but it would help a lot if somebody
> >> could
> >> document (briefly!) the overall purpose of deferred reads and the
> >> exact
> >> purposes
> On Thu, 2008-09-25 at 13:43 -0600, Alex Rousskov wrote:
>>
>> I can probably fix it myself, but it would help a lot if somebody
>> could
>> document (briefly!) the overall purpose of deferred reads and the
>> exact
>> purposes of these two nearly identical methods (one brief description
>> for ea
On Thu, 2008-09-25 at 17:42 -0600, Alex Rousskov wrote:
> Is the order in which reads are "done" or "kicked" important? Does it
> have to be FIFO?
FIFO was the behaviour of the code before I refactored into a manager
object; I suspect FIFO prevents starvation of individual fds, so it
shouldn't b
On Fri, 2008-09-26 at 09:24 +1000, Robert Collins wrote:
> On Thu, 2008-09-25 at 13:43 -0600, Alex Rousskov wrote:
> >
> > I can probably fix it myself, but it would help a lot if somebody
> > could
> > document (briefly!) the overall purpose of deferred reads and the
> > exact
> > purposes of the
On Thu, 2008-09-25 at 13:43 -0600, Alex Rousskov wrote:
>
> I can probably fix it myself, but it would help a lot if somebody
> could
> document (briefly!) the overall purpose of deferred reads and the
> exact
> purposes of these two nearly identical methods (one brief description
> for each, plea
On Thu, 2008-09-25 at 11:32 -0600, Alex Rousskov wrote:
> On Thu, 2008-09-25 at 14:08 +0800, Adrian Chadd wrote:
> > 2008/9/25 Alex Rousskov <[EMAIL PROTECTED]>:
> >
> > > The DescriptorSet class has O(1) complexity for search, insertion,
> > > and deletion. It uses about 2*sizeof(int)*MaxFD bytes
