Add more pdf codec fuzzers.

Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659, 616842, 616841, 616839
Committed: https://crrev.com/7751e5d2a12f07c7538230b6654b1339b7ee2742
Cr-Commit-Position: refs/heads/master@{#398811}

[Oliver Chang, 2016-06-09 03:56:36]

Description was changed from

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
==========

to

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659,616842,616841,616839
==========

[Oliver Chang, 2016-06-09 03:57:39]

ptal. this adds the rest of the fuzzers that Dan implemented.

[Lei Zhang, 2016-06-09 05:20:49]

lgtm

[Oliver Chang, 2016-06-09 07:07:32]

The CQ bit was checked by ochang@chromium.org

[commit-bot: I haz the power, 2016-06-09 07:07:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2043313003/40001

[commit-bot: I haz the power, 2016-06-09 07:49:50]

Description was changed from

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659,616842,616841,616839
==========

to

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659,616842,616841,616839
==========

[commit-bot: I haz the power, 2016-06-09 07:49:51]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-06-09 07:51:38]

Description was changed from

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659,616842,616841,616839
==========

to

==========
Add more pdf codec fuzzers.

R=thestig@chromium.org,dsinclair@chromium.org
BUG=617659,616842,616841,616839

Committed: https://crrev.com/7751e5d2a12f07c7538230b6654b1339b7ee2742
Cr-Commit-Position: refs/heads/master@{#398811}
==========

[commit-bot: I haz the power, 2016-06-09 07:51:39]

Patchset 3 (id:??) landed as
https://crrev.com/7751e5d2a12f07c7538230b6654b1339b7ee2742
Cr-Commit-Position: refs/heads/master@{#398811}

[dsinclair, 2016-06-09 12:59:12]

On 2016/06/09 03:57:39, Oliver Chang wrote:
> ptal. this adds the rest of the fuzzers that Dan implemented.


I was originally holding off landing this change until I could see the results
from the png fuzzer. Looking at what has been found, I'm not sure if it's doing
what we want it to do. It appears the PNG fuzzer is finding issues in the GIF
code, so I'm thinking XFA is ignoring the provided image type flag in some cases
and doing what it wants.

I have a feeling, these are all going to end up testing the same code paths
since they all start with the same test corpii.

[Oliver Chang, 2016-06-09 15:21:32]

On 2016/06/09 12:59:12, dsinclair wrote:
> On 2016/06/09 03:57:39, Oliver Chang wrote:
> > ptal. this adds the rest of the fuzzers that Dan implemented.
> 
> 
> I was originally holding off landing this change until I could see the results
> from the png fuzzer. Looking at what has been found, I'm not sure if it's
doing
> what we want it to do. It appears the PNG fuzzer is finding issues in the GIF
> code, so I'm thinking XFA is ignoring the provided image type flag in some
cases
> and doing what it wants.
> 
> I have a feeling, these are all going to end up testing the same code paths
> since they all start with the same test corpii.

Ah, I haven't looked at the png one too closely yet, and we can always revert.

Either way, we'll want to have different instances running with different
corpora, which I'll upload today. I'll also give you access to PDF security bugs
on CF.

[Oliver Chang, 2016-06-09 15:21:35]

On 2016/06/09 12:59:12, dsinclair wrote:
> On 2016/06/09 03:57:39, Oliver Chang wrote:
> > ptal. this adds the rest of the fuzzers that Dan implemented.
> 
> 
> I was originally holding off landing this change until I could see the results
> from the png fuzzer. Looking at what has been found, I'm not sure if it's
doing
> what we want it to do. It appears the PNG fuzzer is finding issues in the GIF
> code, so I'm thinking XFA is ignoring the provided image type flag in some
cases
> and doing what it wants.
> 
> I have a feeling, these are all going to end up testing the same code paths
> since they all start with the same test corpii.

Ah, I haven't looked at the png one too closely yet, and we can always revert.

Either way, we'll want to have different instances running with different
corpora, which I'll upload today. I'll also give you access to PDF security bugs
on CF.