On 3/23/06, William A. Rowe, Jr. <wrowe@rowe-clan.net> wrote:
> I've refactored the read_with_timeout() code to account for observations
> about both WAIT_ABANDONED and the short race between ReadFile, WaitFor...
> CancelIo and GetCompletion.
>
> It passes regressions which isn't saying much, since a number of edge cases
> in the code are fixed which never appeared in the regression testall framework.
> I need 1) feedback, and 2) opinions...
>
> [ ] code's ready to be adopted as our 'write_with_timeout' logic
> [ ] code should be backported to 1.2.x now
> [ ] code should be backported to 0.9.x now
> [ ] code should be rolled out in the 0.9.11 release already
Considering the lack of regression tests to catch problems with this
code, I think it's rather premature to stick it in 0.9.11.
I've looked over the changes and they seem reasonable to me, but I
don't understand win32 APIs enough to make that a resounding "wow this
is great" kind of response.
-garrett
|