[NaCl SDK] nacl_io: Fix inet_ntop to be consistent with glibc.

[NaCl SDK] nacl_io: Fix inet_ntop to be consistent with glibc.

Update inet_ntop to output ipv4 address for ::A.B.C.D as
well as ::ffff:A:B:C:D so that its consistent the glibc's
implementation.

Also, add two missing socket syscalls: inet_addr and
inet_aton.

Also, add test for this case and cleanup test code these
socket utility function.

Alos, rename some, .cc files to .c in nacl_io syscalls.

R=binji@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=247394

[binji, 2014-01-27 19:57:18]

lgtm, w/ nits

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/hstrerror.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/hstrerror.c:26:
sprintf(unknown_msg, "Unknown error in gethostbyname: %d.", err);
snprintf? I know this will never overrun, but it still makes me have to check...

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_addr.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_addr.c:1: // Copyright
2013 The Chromium Authors. All rights reserved.
2014
here and other new files

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c:14: unsigned int p1
= 0, p2 = 0, p3 = 0, p4 = 0;
Where did this implementation come from? Did you write it?

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c:21: in_addr_t
host_order_addr = p1 << 24 | p2 << 16 | p3 << 8 | p4;
parens around <<? Nicer than having to know operator precedence.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntoa.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntoa.c:17: snprintf(addr,
INET_ADDRSTRLEN, "%hhu.%hhu.%hhu.%hhu",
does it matter? GetByte already returns uint8_t. I had to look up hhu to see
what it actually does

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntop.cc (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntop.cc:38: // Convert to
an arry of 8 host order shorts
sp: array

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntop.cc:64: if
(zero_run_start == 0 && (zero_run_end == 6 || (zero_run_end == 5 &&
might be clearer if you moved the condition after && to its own line:

if (zero_run_start == 0 &&
    (zero_run_end == 6 ||
     (zero_run_end == 5 && ...

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
File native_client_sdk/src/tests/nacl_io_test/socket_test.cc (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:27: // When devloping
tests for host resolution function is can be usefull
When developing tests for host resolution it can be useful

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:325: // Fails for
anything other than
other than...

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:331: // Fails if a
single element is > 256
>= 256

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:366: ASSERT_NE(0,
inet_aton("256", &addr));
test larger numbers here and above? The spec says that x.x.y allows y as 16-bit,
and x.y allows y as 24-bit, and x as 32-bit value.

[Sam Clegg, 2014-01-27 23:10:15]

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/hstrerror.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/hstrerror.c:26:
sprintf(unknown_msg, "Unknown error in gethostbyname: %d.", err);
On 2014/01/27 19:57:18, binji wrote:
> snprintf? I know this will never overrun, but it still makes me have to
check...

Done.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c:14: unsigned int p1
= 0, p2 = 0, p3 = 0, p4 = 0;
On 2014/01/27 19:57:18, binji wrote:
> Where did this implementation come from? Did you write it?

Yup.  I wanted to use inet_pton() but that doesn't work except when there are 4
dotted quads.  inet_aton explictly supports the X, X.Y and X.Y.Z cases :(

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_aton.c:21: in_addr_t
host_order_addr = p1 << 24 | p2 << 16 | p3 << 8 | p4;
On 2014/01/27 19:57:18, binji wrote:
> parens around <<? Nicer than having to know operator precedence.

Done.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntoa.c (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntoa.c:17: snprintf(addr,
INET_ADDRSTRLEN, "%hhu.%hhu.%hhu.%hhu",
On 2014/01/27 19:57:18, binji wrote:
> does it matter? GetByte already returns uint8_t. I had to look up hhu to see
> what it actually does

Done.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
File native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntop.cc (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/li...
native_client_sdk/src/libraries/nacl_io/syscalls/inet_ntop.cc:64: if
(zero_run_start == 0 && (zero_run_end == 6 || (zero_run_end == 5 &&
On 2014/01/27 19:57:18, binji wrote:
> might be clearer if you moved the condition after && to its own line:
> 
> if (zero_run_start == 0 &&
>     (zero_run_end == 6 ||
>      (zero_run_end == 5 && ...

Done.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
File native_client_sdk/src/tests/nacl_io_test/socket_test.cc (right):

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:27: // When devloping
tests for host resolution function is can be usefull
On 2014/01/27 19:57:18, binji wrote:
> When developing tests for host resolution it can be useful

Hmm, this should not be part of this CL.

Removed.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:325: // Fails for
anything other than
On 2014/01/27 19:57:18, binji wrote:
> other than...

Done.

https://codereview.chromium.org/139283003/diff/40001/native_client_sdk/src/te...
native_client_sdk/src/tests/nacl_io_test/socket_test.cc:366: ASSERT_NE(0,
inet_aton("256", &addr));
On 2014/01/27 19:57:18, binji wrote:
> test larger numbers here and above? The spec says that x.x.y allows y as
16-bit,
> and x.y allows y as 24-bit, and x as 32-bit value.

Done.

[Sam Clegg, 2014-01-28 02:33:28]

Committed patchset #2 manually as r247394 (presubmit successful).