Reserve space before building potentially large vectors in safe-browsing.

Reserve space before building potentially large vectors in safe-browsing.

In cases where we can approximate the vector size in advance,
reserve() that much space rather than building organically.

BUG=97102
TEST=Reduction in crash rate for that bug.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=102435

[Scott Hess - ex-Googler, 2011-09-22 20:39:41]

The change in PrefixSetFromAddPrefixes() is the main one, it's almost certain
that it's doing like a thousand cases of "Increase by a
large-but-not-large-enough amount" over and over again, and it's easy to derive
the exact goal size.  The whitelist change is probably overall irrelevant, but
its along the same lines.  The two PrefixSet changes likewise.

[noelutz, 2011-09-22 21:14:09]

lgtm

http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
File chrome/browser/safe_browsing/safe_browsing_database.cc (right):

http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
chrome/browser/safe_browsing/safe_browsing_database.cc:1189:
std::vector<SBAddFullHash> pending_add_hashes;
What about this vector?  Were you thinking that this one was always pretty
small?

[Scott Hess - ex-Googler, 2011-09-22 21:23:09]

thanks!

http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
File chrome/browser/safe_browsing/safe_browsing_database.cc (right):

http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
chrome/browser/safe_browsing/safe_browsing_database.cc:1189:
std::vector<SBAddFullHash> pending_add_hashes;
On 2011/09/22 21:14:09, noelutz wrote:
> What about this vector?  Were you thinking that this one was always pretty
> small?

Yeah, I looked at that one and figured that since it's unlikely to ever have
thousands of entries, it was probably not justified.

Looking at it again, it also seems reasonable to me to assume that
std::vector::insert() has enough information to set the size directly.

[noelutz, 2011-09-22 21:30:57]

lgtm

On Thu, Sep 22, 2011 at 2:23 PM,  <shess@chromium.org> wrote:
> thanks!
>
>
>
http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
> File chrome/browser/safe_browsing/safe_browsing_database.cc (right):
>
>
http://codereview.chromium.org/8005001/diff/1/chrome/browser/safe_browsing/sa...
> chrome/browser/safe_browsing/safe_browsing_database.cc:1189:
> std::vector<SBAddFullHash> pending_add_hashes;
> On 2011/09/22 21:14:09, noelutz wrote:
>>
>> What about this vector?  Were you thinking that this one was always
>
> pretty
>>
>> small?
>
> Yeah, I looked at that one and figured that since it's unlikely to ever
> have thousands of entries, it was probably not justified.
>
> Looking at it again, it also seems reasonable to me to assume that
> std::vector::insert() has enough information to set the size directly.
>
> http://codereview.chromium.org/8005001/
>



-- 
-|||--

[commit-bot: I haz the power, 2011-09-23 01:07:00]

CQ is trying tha patch. Follow status at
https://chromium-status.appspot.com/cq/shess@chromium.org/8005001/1

[commit-bot: I haz the power, 2011-09-23 03:27:57]

Change committed as 102435