Use ARRAYSIZE_UNSAFE when using types defined in a function

Use ARRAYSIZE_UNSAFE when using types defined in a function

Small followup from https://codereview.chromium.org/423333002

This allows the test to compile on a greater range of toolchains (some
of which break when using types defined in a function with arraysize).

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=289051

[Mostyn Bramley-Moore, 2014-08-11 12:21:49]

@wtc: please take a look.

[wtc, 2014-08-11 20:59:17]

Review comments on patch set 1:

Thank you for the CL. I am wondering if the change I suggested below will work.

https://codereview.chromium.org/456403002/diff/1/net/quic/crypto/crypto_utils...
File net/quic/crypto/crypto_utils_test.cc (right):

https://codereview.chromium.org/456403002/diff/1/net/quic/crypto/crypto_utils...
net/quic/crypto/crypto_utils_test.cc:49: const struct TestVector {

This struct is named "TestVector". Why does a compiler consider it as anonymous?

Does it work if we define the struct type separately as follows?

  struct TestVector {
    // Input (strings of hexadecimal digits):
    const char* subkey_secret;
    const char* label;
    const char* context;
    size_t result_len;

    // Expected output (string of hexadecimal digits):
    const char* expected; // Null if it should fail.
  };
  const TestVector  test_vector[] = {
    ...
  };

[Mostyn Bramley-Moore, 2014-08-11 21:59:20]

On 2014/08/11 20:59:17, wtc wrote:
> Review comments on patch set 1:
> 
> Thank you for the CL. I am wondering if the change I suggested below will
work.
> 
>
https://codereview.chromium.org/456403002/diff/1/net/quic/crypto/crypto_utils...
> File net/quic/crypto/crypto_utils_test.cc (right):
> 
>
https://codereview.chromium.org/456403002/diff/1/net/quic/crypto/crypto_utils...
> net/quic/crypto/crypto_utils_test.cc:49: const struct TestVector {
> 
> This struct is named "TestVector". Why does a compiler consider it as
anonymous?

After rereading the comments in base/macros.h I realise that the problem is
really that the compiler doesn't like the fact that the type is defined in a
function: "One caveat is that arraysize() doesn't accept any array of an
anonymous type or a type defined inside a function".  I will adjust the CL
description.  

> Does it work if we define the struct type separately as follows?
> 
>   struct TestVector {
>     // Input (strings of hexadecimal digits):
>     const char* subkey_secret;
>     const char* label;
>     const char* context;
>     size_t result_len;
> 
>     // Expected output (string of hexadecimal digits):
>     const char* expected; // Null if it should fail.
>   };
>   const TestVector  test_vector[] = {
>     ...
>   };

Yes, if I move the definition of the struct outside the test function.  So I
have done this for both tests with arraysize or ARRAYSIZE_UNSAFE in this file.

[wtc, 2014-08-12 14:15:21]

Patch set 1 LGTM.

Mostyn: thank you for the explanation. Now that I know what
the problem is, I prefer your original solution because
defining the structs close to where they are used makes the
code easier to understand. So, if you don't mind, I think
we should go with patch set 1. Please remember to update the
CL's description before you commit it. Thanks.

[Mostyn Bramley-Moore, 2014-08-12 14:20:03]

On 2014/08/12 14:15:21, wtc wrote:
> Patch set 1 LGTM.
> 
> Mostyn: thank you for the explanation. Now that I know what
> the problem is, I prefer your original solution because
> defining the structs close to where they are used makes the
> code easier to understand. So, if you don't mind, I think
> we should go with patch set 1. Please remember to update the
> CL's description before you commit it. Thanks.

Done, thanks for the review.

[Mostyn Bramley-Moore, 2014-08-12 14:20:09]

The CQ bit was checked by mostynb@opera.com

[commit-bot: I haz the power, 2014-08-12 14:22:41]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/mostynb@opera.com/456403002/40001

[commit-bot: I haz the power, 2014-08-12 15:52:30]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  android_dbg_triggered_tests on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/android_dbg_tri...)

[commit-bot: I haz the power, 2014-08-12 19:39:27]

Change committed as 289051