Let RedirectToFileResourceHandler resize its internal buffer.

Let RedirectToFileResourceHandler resize its internal buffer.

This can decrease NaCl's total warm load time by ~15% for larger files.
The algorithm for resizing the buffer matches the one used by
AsyncResourceHandler.

BUG= none
TEST= none


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=148664

[Nick Bray (chromium), 2012-07-25 23:08:29]

darin: listing you are reviewer because you are OWNER and blamed for most of
this file.  Feel free to kick it to someone else if your review queue is backed
up.

[darin (slow to review), 2012-07-25 23:20:37]

I wonder about how we might share this code (this algorithm) between
AsyncResourceHandler and RedirectToFileResourceHandler.  It seems a tad
unfortunate to replicate the code.

[Nick Bray (chromium), 2012-07-25 23:36:43]

On 2012/07/25 23:20:37, darin wrote:
> I wonder about how we might share this code (this algorithm) between
> AsyncResourceHandler and RedirectToFileResourceHandler.  It seems a tad
> unfortunate to replicate the code.

I thought about this, but didn't see a good way to do it.  The algorithm part is
very simple yet the actual implementations are wildly divergent.  The ways I've
thought about abstracting out the algorithm (a common base class, for example)
would add complexity that outweighed the sharing.  The constants could be
shared, but again it wasn't apparent that sharing them gained much, particularly
as they seemed to be arbitrarily chosen.

[darin (slow to review), 2012-07-26 18:51:59]

https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/render...
File content/browser/renderer_host/redirect_to_file_resource_handler.cc (right):

https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/render...
content/browser/renderer_host/redirect_to_file_resource_handler.cc:115: if
(buf_->capacity() == bytes_read) {
I had just imagined a BufferSizeAlgorithm class with methods like:

size_t GetNextAllocationSize();
void UpdateNextAllocationSize(size_t bytes_used_in_last_allocation);

[darin (slow to review), 2012-07-26 18:52:56]

On Thu, Jul 26, 2012 at 11:51 AM, <darin@chromium.org> wrote:

>
> https://chromiumcodereview.**appspot.com/10828024/diff/1/**
>
content/browser/renderer_host/**redirect_to_file_resource_**handler.cc<https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/renderer_host/redirect_to_file_resource_handler.cc>
> File content/browser/renderer_host/**redirect_to_file_resource_**
> handler.cc
> (right):
>
> https://chromiumcodereview.**appspot.com/10828024/diff/1/**
> content/browser/renderer_host/**redirect_to_file_resource_**
>
handler.cc#newcode115<https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/renderer_host/redirect_to_file_resource_handler.cc#newcode115>
> content/browser/renderer_host/**redirect_to_file_resource_**
> handler.cc:115:
>
> if (buf_->capacity() == bytes_read) {
> I had just imagined a BufferSizeAlgorithm class with methods like:
>

s/BufferSizeAlgorithm/BufferSizeStrategy/  <-- probably better


>
> size_t GetNextAllocationSize();
> void UpdateNextAllocationSize(size_**t bytes_used_in_last_allocation)**;
>
>
https://chromiumcodereview.**appspot.com/10828024/<https://chromiumcodereview...
>

[darin (slow to review), 2012-07-26 19:12:01]

As discussed in person, LGTM for this CL.

On Thu, Jul 26, 2012 at 11:52 AM, Darin Fisher <darin@chromium.org> wrote:

> On Thu, Jul 26, 2012 at 11:51 AM, <darin@chromium.org> wrote:
>
>>
>> https://chromiumcodereview.**appspot.com/10828024/diff/1/**
>>
content/browser/renderer_host/**redirect_to_file_resource_**handler.cc<https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/renderer_host/redirect_to_file_resource_handler.cc>
>> File content/browser/renderer_host/**redirect_to_file_resource_**
>> handler.cc
>> (right):
>>
>> https://chromiumcodereview.**appspot.com/10828024/diff/1/**
>> content/browser/renderer_host/**redirect_to_file_resource_**
>>
handler.cc#newcode115<https://chromiumcodereview.appspot.com/10828024/diff/1/content/browser/renderer_host/redirect_to_file_resource_handler.cc#newcode115>
>> content/browser/renderer_host/**redirect_to_file_resource_**
>> handler.cc:115:
>>
>> if (buf_->capacity() == bytes_read) {
>> I had just imagined a BufferSizeAlgorithm class with methods like:
>>
>
> s/BufferSizeAlgorithm/BufferSizeStrategy/  <-- probably better
>
>
>>
>> size_t GetNextAllocationSize();
>> void UpdateNextAllocationSize(size_**t bytes_used_in_last_allocation)**;
>>
>>
https://chromiumcodereview.**appspot.com/10828024/<https://chromiumcodereview...
>>
>
>

[commit-bot: I haz the power, 2012-07-26 20:47:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ncbray@chromium.org/10828024/1

[commit-bot: I haz the power, 2012-07-26 23:57:15]

Change committed as 148664