Cleanup some PPAPI proxy code.

Cleanup some PPAPI proxy code.

Check for potential overflows in PDFResource::SearchString().

Committed: https://crrev.com/c1f16bd8639372d9e45cdac4e384a550d561af2e
Cr-Commit-Position: refs/heads/master@{#330675}

[Lei Zhang, 2015-05-19 22:20:01]

thestig@chromium.org changed reviewers:
+ raymes@chromium.org

[raymes, 2015-05-20 00:26:48]

lgtm

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
File ppapi/proxy/pdf_resource.cc (right):

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
ppapi/proxy/pdf_resource.cc:104: std::numeric_limits<size_t>::max() /
sizeof(PP_PrivateFindResult)) {
Would it be simpler just to have some large constant maximum?

kMaxNumSearchResults = 1000000;

[Lei Zhang, 2015-05-20 00:30:48]

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
File ppapi/proxy/pdf_resource.cc (right):

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
ppapi/proxy/pdf_resource.cc:104: std::numeric_limits<size_t>::max() /
sizeof(PP_PrivateFindResult)) {
On 2015/05/20 00:26:48, raymes wrote:
> Would it be simpler just to have some large constant maximum?
> 
> kMaxNumSearchResults = 1000000;

I can replace std::numeric_limits<size_t>::max() with SIZE_MAX. That'll make it
less verbose.

[Lei Zhang, 2015-05-20 00:46:34]

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
File ppapi/proxy/pdf_resource.cc (right):

https://codereview.chromium.org/1147883002/diff/20001/ppapi/proxy/pdf_resourc...
ppapi/proxy/pdf_resource.cc:104: std::numeric_limits<size_t>::max() /
sizeof(PP_PrivateFindResult)) {
On 2015/05/20 00:30:48, Lei Zhang wrote:
> On 2015/05/20 00:26:48, raymes wrote:
> > Would it be simpler just to have some large constant maximum?
> > 
> > kMaxNumSearchResults = 1000000;
> 
> I can replace std::numeric_limits<size_t>::max() with SIZE_MAX. That'll make
it
> less verbose.

Done in patch set 3. But in general, the security folks do these kinds of
checks, or use base/numerics/safe_conversions.h and friends all the time, rather
than set some arbitrary limit. The user's machine may run out of memory, but
that's not my concern.

[raymes, 2015-05-20 00:48:52]

That's ok, I'll leave it up to you :) I was mainly just thinking from a code
readability standpoint and reasoning that placing a cap on the number of the
results wouldn't be too unreasonable. But I can see arguments against this too
:)

lgtm

[Lei Zhang, 2015-05-20 00:53:26]

The CQ bit was checked by thestig@chromium.org

[commit-bot: I haz the power, 2015-05-20 00:58:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1147883002/40001

[commit-bot: I haz the power, 2015-05-20 02:01:38]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-05-20 02:04:31]

Patchset 3 (id:??) landed as
https://crrev.com/c1f16bd8639372d9e45cdac4e384a550d561af2e
Cr-Commit-Position: refs/heads/master@{#330675}