Do an in-place replacement of SkRandom with Fuzz for FilterFuzz

Do an in-place replacement of SkRandom with Fuzz for FilterFuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1710183002

Committed: https://skia.googlesource.com/skia/+/4319593988db5796023d9f5f34a8ed285c2097dd

[kjlubick, 2016-02-18 20:47:35]

Description was changed from

==========
Do an in-place replacement of SkRandom with Fuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
==========

to

==========
Do an in-place replacement of SkRandom with Fuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[kjlubick, 2016-02-18 20:52:57]

kjlubick@google.com changed reviewers:
+ jcgregorio@google.com, reed@google.com

[kjlubick, 2016-02-18 20:52:58]

This is a naive approach to hooking our new fuzzer up to this code.  A more
powerful fuzzing approach would replace make_number with a direct feed to the
fuzzer. I tried this first, but the fuzzer found a crash almost immediately, so
perhaps I did it wrong, or the code is too fragile.

A third possible approach is available if image filters can be serialized to a
binary format on disk.  Then, we can give the fuzzer some good filters and have
it go from there.

What do you think?

[kjlubick, 2016-02-18 20:53:13]

Description was changed from

==========
Do an in-place replacement of SkRandom with Fuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
Do an in-place replacement of SkRandom with Fuzz for FilterFuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[kjlubick, 2016-02-19 13:57:03]

On 2016/02/18 at 20:52:58, kjlubick wrote:
> This is a naive approach to hooking our new fuzzer up to this code.  A more
powerful fuzzing approach would replace make_number with a direct feed to the
fuzzer. I tried this first, but the fuzzer found a crash almost immediately, so
perhaps I did it wrong, or the code is too fragile.
> 
> A third possible approach is available if image filters can be serialized to a
binary format on disk.  Then, we can give the fuzzer some good filters and have
it go from there.
> 
> What do you think?

I ran this overnight on my desktop and it found a few crashes.  Want to take a
look?

[kjlubick, 2016-04-05 19:15:06]

The CQ bit was checked by kjlubick@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-05 19:15:14]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1710183002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1710183002/60001

[commit-bot: I haz the power, 2016-04-05 19:30:44]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-04-05 19:30:45]

Dry run: This issue passed the CQ dry run.

[jcgregorio, 2016-04-05 19:31:14]

https://codereview.chromium.org/1710183002/diff/60001/fuzz/fuzz.cpp
File fuzz/fuzz.cpp (right):

https://codereview.chromium.org/1710183002/diff/60001/fuzz/fuzz.cpp#newcode402
fuzz/fuzz.cpp:402: unsigned int floatint = 0x3f800000 | (this->nextU() >> 9);
Maybe just me, but I think this deserves a comment.

[kjlubick, 2016-04-05 19:34:49]

https://codereview.chromium.org/1710183002/diff/60001/fuzz/fuzz.cpp
File fuzz/fuzz.cpp (right):

https://codereview.chromium.org/1710183002/diff/60001/fuzz/fuzz.cpp#newcode402
fuzz/fuzz.cpp:402: unsigned int floatint = 0x3f800000 | (this->nextU() >> 9);
On 2016/04/05 at 19:31:14, jcgregorio wrote:
> Maybe just me, but I think this deserves a comment.

Done.  I just copy-pasted it from SkRandom

[jcgregorio, 2016-04-05 19:37:44]

lgtm

[kjlubick, 2016-04-05 19:38:20]

The CQ bit was checked by kjlubick@google.com

[commit-bot: I haz the power, 2016-04-05 19:38:26]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1710183002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1710183002/80001

[commit-bot: I haz the power, 2016-04-05 19:48:49]

Description was changed from

==========
Do an in-place replacement of SkRandom with Fuzz for FilterFuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
Do an in-place replacement of SkRandom with Fuzz for FilterFuzz

This feels rather clunky, because we aren't using the full potential of the
fuzzer, but it works, it seems.

BUG=skia:4969
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/4319593988db5796023d9f5f34a8ed285c2097dd
==========

[commit-bot: I haz the power, 2016-04-05 19:48:51]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/4319593988db5796023d9f5f34a8ed285c2097dd