Convert the path fuzzer test to a unit test

Convert the path fuzzer test to a unit test

This salvages the interesting bits of svg/dom/fuzz-path-parser.html into the
SVGPathParserTest.Simple unit test.

BUG=539385
Committed: https://crrev.com/cffa5e8e66d1379b530dba77bb67d95a9aaf8bc4
Cr-Commit-Position: refs/heads/master@{#355282}

[fs, 2015-10-20 17:18:17]

fs@opera.com changed reviewers:
+ pdr@chromium.org

[fs, 2015-10-20 17:18:18]

So, there were some options...

1) Remove the test.
2) Remove the "output of the parsed result"-part of the test.
3) Convert to unit test, preserving behavior.
4) Convert to unit test, use different random sequence.

The CL in its current form is (3), but I'd consider any other of the above to
still be on the table. I guess it depends a bit how useful we consider this test
to be (not overly so I think.)

Please let me know what you think.

[pdr., 2015-10-20 18:31:43]

On 2015/10/20 at 17:18:18, fs wrote:
> So, there were some options...
> 
> 1) Remove the test.
> 2) Remove the "output of the parsed result"-part of the test.
> 3) Convert to unit test, preserving behavior.
> 4) Convert to unit test, use different random sequence.
> 
> The CL in its current form is (3), but I'd consider any other of the above to
still be on the table. I guess it depends a bit how useful we consider this test
to be (not overly so I think.)
> 
> Please let me know what you think.

I'm not a huge fan of these fuzzy tests :(

Is there anything in here that isn't covered by your other tests? If so, maybe
we can just distill this test down to a few good test cases? If not, I'm more in
favor of #1.

[fs, 2015-10-20 18:44:36]

On 2015/10/20 at 18:31:43, pdr wrote:
> On 2015/10/20 at 17:18:18, fs wrote:
> > So, there were some options...
> > 
> > 1) Remove the test.
> > 2) Remove the "output of the parsed result"-part of the test.
> > 3) Convert to unit test, preserving behavior.
> > 4) Convert to unit test, use different random sequence.
> > 
> > The CL in its current form is (3), but I'd consider any other of the above
to still be on the table. I guess it depends a bit how useful we consider this
test to be (not overly so I think.)
> > 
> > Please let me know what you think.
> 
> I'm not a huge fan of these fuzzy tests :(

You and me both then...
 
> Is there anything in here that isn't covered by your other tests? If so, maybe
we can just distill this test down to a few good test cases? If not, I'm more in
favor of #1.

Can't be ruled out really, and I'd need to do some archeology to find out I
think. I'll see what I can dig up - salvage that - and then execute #1.

[fs, 2015-10-20 19:24:39]

On 2015/10/20 at 18:44:36, fs wrote:
> On 2015/10/20 at 18:31:43, pdr wrote:
> > On 2015/10/20 at 17:18:18, fs wrote:
> > > So, there were some options...
> > > 
> > > 1) Remove the test.
> > > 2) Remove the "output of the parsed result"-part of the test.
> > > 3) Convert to unit test, preserving behavior.
> > > 4) Convert to unit test, use different random sequence.
> > > 
> > > The CL in its current form is (3), but I'd consider any other of the above
to still be on the table. I guess it depends a bit how useful we consider this
test to be (not overly so I think.)
> > > 
> > > Please let me know what you think.
> > 
> > I'm not a huge fan of these fuzzy tests :(
> 
> You and me both then...
>  
> > Is there anything in here that isn't covered by your other tests? If so,
maybe we can just distill this test down to a few good test cases? If not, I'm
more in favor of #1.
> 
> Can't be ruled out really, and I'd need to do some archeology to find out I
think. I'll see what I can dig up - salvage that - and then execute #1.

*removes fedora and shelves the whip again*

Looks, like this test was added in http://trac.webkit.org/changeset/20375 -
hopefully it should be possible to boil that down to something more to the
point.

[fs, 2015-10-21 09:19:43]

Description was changed from

==========
Convert the path fuzzer test to a unit test

This turns the svg/dom/fuzz-path-parser.html into a unit test.

BUG=539385
==========

to

==========
Convert the path fuzzer test to a unit test

This salvages the interesting bits of svg/dom/fuzz-path-parser.html into the
SVGPathParserTest.Simple unit test.

BUG=539385
==========

[fs, 2015-10-21 10:16:15]

On 2015/10/20 at 19:24:39, fs wrote:
> On 2015/10/20 at 18:44:36, fs wrote:
> > On 2015/10/20 at 18:31:43, pdr wrote:
> > > On 2015/10/20 at 17:18:18, fs wrote:
> > > > So, there were some options...
> > > > 
> > > > 1) Remove the test.
> > > > 2) Remove the "output of the parsed result"-part of the test.
> > > > 3) Convert to unit test, preserving behavior.
> > > > 4) Convert to unit test, use different random sequence.
> > > > 
> > > > The CL in its current form is (3), but I'd consider any other of the
above to still be on the table. I guess it depends a bit how useful we consider
this test to be (not overly so I think.)
> > > > 
> > > > Please let me know what you think.
> > > 
> > > I'm not a huge fan of these fuzzy tests :(
> > 
> > You and me both then...
> >  
> > > Is there anything in here that isn't covered by your other tests? If so,
maybe we can just distill this test down to a few good test cases? If not, I'm
more in favor of #1.
> > 
> > Can't be ruled out really, and I'd need to do some archeology to find out I
think. I'll see what I can dig up - salvage that - and then execute #1.
> 
> *removes fedora and shelves the whip again*
> 
> Looks, like this test was added in http://trac.webkit.org/changeset/20375 -
hopefully it should be possible to boil that down to something more to the
point.

Salvage completed, PTAL.

[pdr., 2015-10-21 13:23:01]

On 2015/10/21 at 10:16:15, fs wrote:
> On 2015/10/20 at 19:24:39, fs wrote:
> > On 2015/10/20 at 18:44:36, fs wrote:
> > > On 2015/10/20 at 18:31:43, pdr wrote:
> > > > On 2015/10/20 at 17:18:18, fs wrote:
> > > > > So, there were some options...
> > > > > 
> > > > > 1) Remove the test.
> > > > > 2) Remove the "output of the parsed result"-part of the test.
> > > > > 3) Convert to unit test, preserving behavior.
> > > > > 4) Convert to unit test, use different random sequence.
> > > > > 
> > > > > The CL in its current form is (3), but I'd consider any other of the
above to still be on the table. I guess it depends a bit how useful we consider
this test to be (not overly so I think.)
> > > > > 
> > > > > Please let me know what you think.
> > > > 
> > > > I'm not a huge fan of these fuzzy tests :(
> > > 
> > > You and me both then...
> > >  
> > > > Is there anything in here that isn't covered by your other tests? If so,
maybe we can just distill this test down to a few good test cases? If not, I'm
more in favor of #1.
> > > 
> > > Can't be ruled out really, and I'd need to do some archeology to find out
I think. I'll see what I can dig up - salvage that - and then execute #1.
> > 
> > *removes fedora and shelves the whip again*
> > 
> > Looks, like this test was added in http://trac.webkit.org/changeset/20375 -
hopefully it should be possible to boil that down to something more to the
point.
> 
> Salvage completed, PTAL.

LGTM

[fs, 2015-10-21 13:27:00]

The CQ bit was checked by fs@opera.com

[commit-bot: I haz the power, 2015-10-21 13:27:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1407133007/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1407133007/40001

[commit-bot: I haz the power, 2015-10-21 13:31:26]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-10-21 13:31:58]

Patchset 3 (id:??) landed as
https://crrev.com/cffa5e8e66d1379b530dba77bb67d95a9aaf8bc4
Cr-Commit-Position: refs/heads/master@{#355282}