Increase kMaxHostsSize to 32M

Increase kMaxHostsSize to 32M

Now that hosts file parsing is faster, increase the maximum file size
Chrome will read from 64k to 32M.

BUG=107810
TEST=DnsHostsTest continues to pass

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=210957

[Deprecated (see juliatuttle), 2013-07-09 18:43:34]

PTAL, szym.

[mmenke, 2013-07-09 23:36:44]

On 2013/07/09 18:43:34, ttuttle wrote:
> PTAL, szym.

This may be too big.  Suppose each entry takes up 20 bytes, on average.  That
would be 640 MB RAM, which is a lot, particularly with a 32-bit address space. 
Don't suppose either of you knows if Chrome can use 4GB of address space on
Windows, or just 2GB?  Believe there's a flag in each windows binary indicating
if the program is "large address space aware" (Or something like that).

[szym, 2013-07-09 23:38:32]

On 2013/07/09 23:36:44, mmenke wrote:
> On 2013/07/09 18:43:34, ttuttle wrote:
> > PTAL, szym.
> 
> This may be too big.  Suppose each entry takes up 20 bytes, on average.  That
> would be 640 MB RAM, which is a lot, particularly with a 32-bit address space.

> Don't suppose either of you knows if Chrome can use 4GB of address space on
> Windows, or just 2GB?  Believe there's a flag in each windows binary
indicating
> if the program is "large address space aware" (Or something like that).

kMaxHostSize is in bytes, not entries. I suppose a comment would help.

The CL, LGTM, but pay close attention to the parse duration histograms.

[mmenke, 2013-07-10 01:30:57]

On 2013/07/09 23:38:32, szym wrote:
> On 2013/07/09 23:36:44, mmenke wrote:
> > On 2013/07/09 18:43:34, ttuttle wrote:
> > > PTAL, szym.
> > 
> > This may be too big.  Suppose each entry takes up 20 bytes, on average. 
That
> > would be 640 MB RAM, which is a lot, particularly with a 32-bit address
space.
> 
> > Don't suppose either of you knows if Chrome can use 4GB of address space on
> > Windows, or just 2GB?  Believe there's a flag in each windows binary
> indicating
> > if the program is "large address space aware" (Or something like that).
> 
> kMaxHostSize is in bytes, not entries. I suppose a comment would help.
> 
> The CL, LGTM, but pay close attention to the parse duration histograms.

Or renaming it to kMaxHostsFileSize.

I didn't even look at the code, just saw the description.  Regardless, I
consider the point addressed, and the change LGTM as-is.

[commit-bot: I haz the power, 2013-07-10 18:27:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ttuttle@chromium.org/18302003/6001

[commit-bot: I haz the power, 2013-07-10 23:03:55]

Change committed as 210957