Add a call to MemoryDebug::MarkAsInitialized to the new network code to avoid...

Add a call to MemoryDebug::MarkAsInitialized to the new network code to avoid
false UMR reports from Purify due to asynchronous filling of memory from the 
socket.

BUG=5297
TEST=run test_shell.exe http://localhost:8080/multipart/invalid-image-data-standalone.html
  in Purify, see no UMRs in net code.  Or watch the layout-test Purify
  buildbot running http/tests/invalid-image-data{-standalone}.html

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=7028

[darin (slow to review), 2008-12-10 17:33:29]

LGTM

http://codereview.chromium.org/13331/diff/202/203
File net/base/tcp_client_socket_win.cc (right):

http://codereview.chromium.org/13331/diff/202/203#newcode185
Line 185: // See http://code.google.com/p/chromium/issues/detail?id=5297
Can we use a more abbreviated way of referring to DIT issues?  How about
http://crbug.com/5297 or just bug 5297 since we have b/12345678 for buganizer
references?

[Pam (message me for reviews), 2008-12-10 21:37:50]

On 2008/12/10 17:33:29, darin wrote:
> LGTM
> 
> http://codereview.chromium.org/13331/diff/202/203
> File net/base/tcp_client_socket_win.cc (right):
> 
> http://codereview.chromium.org/13331/diff/202/203#newcode185
> Line 185: // See http://code.google.com/p/chromium/issues/detail?id=5297
> Can we use a more abbreviated way of referring to DIT issues?  How about
> http://crbug.com/5297 or just bug 5297 since we have b/12345678 for buganizer
> references?

My understanding is that we're not supposed to be using crbug.com as the
"canonical URL" for bug links.  See Ian's email of November 25.  I'm fine with
"bug 5297", although it's more work to get there than clicking on a link would
be.

- Pam