Include-what-you-use cleanup for http_cache_transaction.*

Include-what-you-use cleanup for http_cache_transaction.*

Add missing includes to http_cache_transaction.h and
http_cache_transaction.cc.

Remove includes from http_cache_transaction.cc that duplicate those in
http_cache_transaction.h.

Remove includes from http_cache_transaction.cc that are no longer used.

BUG=none
TEST=net_unittests

Committed: https://crrev.com/d4470367e93495c078f7443aecc20dab88d48165
Cr-Commit-Position: refs/heads/master@{#334998}

[Adam Rice, 2015-06-17 10:13:04]

ricea@chromium.org changed reviewers:
+ gavinp@chromium.org

[gavinp, 2015-06-17 13:56:14]

lgtm, although I think you should pay attention to the IOBuffer thing.

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
File net/http/http_cache_transaction.cc (left):

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.cc:28: #include "base/rand_util.h"
!!! Wow.

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
File net/http/http_cache_transaction.h (right):

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.h:10: 
Should we also #include <stddef.h> for size_t?

(in an earlier comment on another review, I think I referred to <stdint.h> for
this?)

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.h:34: class IOBuffer;
We use a scoped_refptr of IOBuffer (read_buf_ on line 462). That makes using a
forward declaration quite fragile; if any constructors or destructors are
generated then we could blow up. Should we just add an include?

(I didn't know this even worked! I spent some time playing with it to convince
myself that clang at least compiles scoped_refptr to a forward declared class,
although it is pretty easy to generate a method that blows it up...)

[Adam Rice, 2015-06-18 03:56:21]

The CQ bit was checked by ricea@chromium.org

[Adam Rice, 2015-06-18 03:56:21]

The patchset sent to the CQ was uploaded after l-g-t-m from gavinp@chromium.org
Link to the patchset: https://codereview.chromium.org/1178013004/#ps40001
(title: "Rabse to fix merge error.")

[commit-bot: I haz the power, 2015-06-18 03:57:40]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1178013004/40001

[commit-bot: I haz the power, 2015-06-18 05:48:09]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-06-18 05:50:17]

Patchset 3 (id:??) landed as
https://crrev.com/d4470367e93495c078f7443aecc20dab88d48165
Cr-Commit-Position: refs/heads/master@{#334998}

[Adam Rice, 2015-06-19 07:00:01]

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
File net/http/http_cache_transaction.cc (left):

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.cc:28: #include "base/rand_util.h"
On 2015/06/17 13:56:14, gavinp wrote:
> !!! Wow.

Yes.

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
File net/http/http_cache_transaction.h (right):

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.h:10: 
On 2015/06/17 13:56:14, gavinp wrote:
> Should we also #include <stddef.h> for size_t?
> 
> (in an earlier comment on another review, I think I referred to <stdint.h> for
> this?)

Done.

This file still uses the deprecated integer types from basictypes.h. Changing
those would be more disruptive, so I left them alone.

https://codereview.chromium.org/1178013004/diff/1/net/http/http_cache_transac...
net/http/http_cache_transaction.h:34: class IOBuffer;
On 2015/06/17 13:56:14, gavinp wrote:
> We use a scoped_refptr of IOBuffer (read_buf_ on line 462). That makes using a
> forward declaration quite fragile; if any constructors or destructors are
> generated then we could blow up. Should we just add an include?
> 
> (I didn't know this even worked! I spent some time playing with it to convince
> myself that clang at least compiles scoped_refptr to a forward declared class,
> although it is pretty easy to generate a method that blows it up...)

No point in unnecessary excitement. I moved the io_buffer.h include to this
file.