Remove NSInputStream used in HTTPTransportMac and use a CFReadStream instead.

Remove NSInputStream used in HTTPTransportMac and use a CFReadStream instead.

NSInputStream requires overriding and implementing private methods in order to
use it with NSURLConnection [1]. It is cleaner to use the private but stable
and open source CFStreamAbstract.h header from CF-Lite to implement a
CFReadStream. Since CFReadStream is toll-free bridged to NSInputStream, the
remainder of the HTTPTransport code can remain unchanged.

[1] http://lists.apple.com/archives/macnetworkprog/2007/May/msg00055.html

BUG=crashpad:15
R=mark@chromium.org

Committed: https://chromium.googlesource.com/crashpad/crashpad/+/f0ee5f0efee651ab82aa854761f107193b3db5de

[Mark Mentovai, 2015-03-11 20:47:42]

https://codereview.chromium.org/993413003/diff/1/third_party/apple_cf/README....
File third_party/apple_cf/README.crashpad (right):

https://codereview.chromium.org/993413003/diff/1/third_party/apple_cf/README....
third_party/apple_cf/README.crashpad:7: License File: APPLE_LICENSE
Should be LICENSE, since that’s what you’re actually checking in. Or rename the
file so that it matches what it’s called upstream.

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac.mm
File util/net/http_transport_mac.mm (right):

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:25: #include
"third_party/apple_cf/CFStreamAbstract.h"
third_party < util

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:41: CFStreamClientContext context = {
Can these both be const?

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:60: CFReadStreamRef read_stream =
CFReadStreamCreate(NULL,
nullptr

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:90: ssize_t bytes_read =
GetStream(info)->GetBytesBuffer(buffer, buffer_length);
Paranoia: if buffer_length == 0, just return 0.

[Robert Sesek, 2015-03-11 20:53:57]

https://codereview.chromium.org/993413003/diff/1/third_party/apple_cf/README....
File third_party/apple_cf/README.crashpad (right):

https://codereview.chromium.org/993413003/diff/1/third_party/apple_cf/README....
third_party/apple_cf/README.crashpad:7: License File: APPLE_LICENSE
On 2015/03/11 20:47:42, Mark Mentovai wrote:
> Should be LICENSE, since that’s what you’re actually checking in. Or rename
the
> file so that it matches what it’s called upstream.

Done.

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac.mm
File util/net/http_transport_mac.mm (right):

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:25: #include
"third_party/apple_cf/CFStreamAbstract.h"
On 2015/03/11 20:47:42, Mark Mentovai wrote:
> third_party < util

Done.

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:41: CFStreamClientContext context = {
On 2015/03/11 20:47:42, Mark Mentovai wrote:
> Can these both be const?

No, only the callbacks.

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:60: CFReadStreamRef read_stream =
CFReadStreamCreate(NULL,
On 2015/03/11 20:47:42, Mark Mentovai wrote:
> nullptr

Done. Force of habit with CF.

https://codereview.chromium.org/993413003/diff/1/util/net/http_transport_mac....
util/net/http_transport_mac.mm:90: ssize_t bytes_read =
GetStream(info)->GetBytesBuffer(buffer, buffer_length);
On 2015/03/11 20:47:42, Mark Mentovai wrote:
> Paranoia: if buffer_length == 0, just return 0.

Done.

[Mark Mentovai, 2015-03-11 20:59:33]

LGTM. I tested exhaustively on 10.6.8, 10.7.5, 10.8.5, 10.9.5, and 10.10.2. All
tests passed and I was able to use the run_with_crashpad/crashpad_handler
combination to upload crash reports to the staging server on all five systems.

[Robert Sesek, 2015-03-11 21:00:04]

Committed patchset #2 (id:20001) manually as
f0ee5f0efee651ab82aa854761f107193b3db5de (presubmit successful).