On Tuesday 26 September 2017 02:28 AM, Ivan Gerasimov wrote:
Thank you Roger for review!
On 9/25/17 11:47 AM, Roger Riggs wrote:
Hi Ivan,
Looks ok to me.
I don't see a reason to skimp on the additional size since it is
allocated and then freed immediately.
The increment might as well be
Hi Ivan,
Looks fine,
Thanks, Roger
On 9/25/2017 4:58 PM, Ivan Gerasimov wrote:
Thank you Roger for review!
On 9/25/17 11:47 AM, Roger Riggs wrote:
Hi Ivan,
Looks ok to me.
I don't see a reason to skimp on the additional size since it is
allocated and then freed immediately.
The incremen
Thank you Roger for review!
On 9/25/17 11:47 AM, Roger Riggs wrote:
Hi Ivan,
Looks ok to me.
I don't see a reason to skimp on the additional size since it is
allocated and then freed immediately.
The increment might as well be as big as the initial size.
Right. Let's use the same value of
Hi Ivan,
Looks ok to me.
I don't see a reason to skimp on the additional size since it is
allocated and then freed immediately.
The increment might as well be as big as the initial size.
I don't see a reason to retry if the buffer gets close to ULONG_MAX; I'd
just break the for loop
and let
Ping.
Please review the proposed change at your convenience.
The fix will greatly reduce the possibility of a need to reallocate the
buffer to retrieve the results (something that the documentation
strongly suggests to avoid), and, if such reallocation still occurs to
be necessary, will incre
Hello!
When retrieving information about network interfaces on Windows we make
up to 2 attempts to call GetAdaptersAddresses().
It was reported that in very rare cases it may not be sufficient, and
even the second attempt can fail with ERROR_BUFFER_OVERFLOW.
I suggest that we follow the rec
