Disable dartc diet parser test.

Disable dartc diet parser test.

This test is causing intermittent redness (and may no longer be of use).

Disabling pending further investigation.

Committed: https://code.google.com/p/dart/source/detail?r=9380

[Brian Wilkerson, 2012-07-03 19:31:50]

As far as I know, these tests are still of use.

Which test is causing redness? All of them?

> Disabling pending further investigation.

Will you be doing the investigation?

[pquitslund, 2012-07-03 20:18:19]

On 2012/07/03 19:31:50, Brian Wilkerson wrote:
> As far as I know, these tests are still of use.
> 
> Which test is causing redness? All of them?

DietParserTest is timing out intermittently.

> 
> > Disabling pending further investigation.
> 
> Will you be doing the investigation?

Right.  I wasn't actually volunteering but someone should own it.  The sense I
got from zundel is that this particular test may no longer be valuable and so
might in fact be a candidate for removal. I was hoping you guys would chime in
to confirm or deny.  Of course, and either way, the fact that it occasionally
times out is interesting and IS telling us something valuable...

In the end, I'm happy not disabling it.  I was just trying to bring some more
stability to our tests which frequently distract us (not to mention others). 
That being said, it would be interesting to know why this is suddenly happening.
 Along those lines, I wonder if it has anything to do with the recent
performance regression reports?  That is, maybe recent parser slow-downs are at
the root of this failure?

[Brian Wilkerson, 2012-07-03 20:37:22]

> DietParserTest is timing out intermittently.

I did some additional digging. It's the method testTiming() that's timing out.
On my Linux desktop it's running in 400 +/- 5 ms. but it's sometimes taking more
than twice that long on the build bots. I don't know why.

> The sense I got from zundel is that this particular test may no longer be
valuable
> and so might in fact be a candidate for removal. I was hoping you guys would
> chime in to confirm or deny.

Unfortunately, I don't know.

> I was just trying to bring some more stability to our tests which frequently
> distract us (not to mention others).

I'm all for stability as long as it isn't hiding problems.

> That being said, it would be interesting to know why this is suddenly
happening.

Agreed.

>  Along those lines, I wonder if it has anything to do with the recent
> performance regression reports?  That is, maybe recent parser slow-downs are
at
> the root of this failure?

That seems likely. Guess it's time to do some profiling.

[zundel, 2012-07-03 20:39:49]

My understanding is that diet parsing is used only for incremental compilation
which is turned off.

On 2012/07/03 20:37:22, Brian Wilkerson wrote:
> > DietParserTest is timing out intermittently.
> 
> I did some additional digging. It's the method testTiming() that's timing out.
> On my Linux desktop it's running in 400 +/- 5 ms. but it's sometimes taking
more
> than twice that long on the build bots. I don't know why.
> 
> > The sense I got from zundel is that this particular test may no longer be
> valuable
> > and so might in fact be a candidate for removal. I was hoping you guys would
> > chime in to confirm or deny.
> 
> Unfortunately, I don't know.
> 
> > I was just trying to bring some more stability to our tests which frequently
> > distract us (not to mention others).
> 
> I'm all for stability as long as it isn't hiding problems.
> 
> > That being said, it would be interesting to know why this is suddenly
> happening.
> 
> Agreed.
> 
> >  Along those lines, I wonder if it has anything to do with the recent
> > performance regression reports?  That is, maybe recent parser slow-downs are
> at
> > the root of this failure?
> 
> That seems likely. Guess it's time to do some profiling.

[Brian Wilkerson, 2012-07-03 21:17:53]

> My understanding is that diet parsing is used only for incremental compilation
> which is turned off.

In which case we might as well disable the tests.

Phil: LGTM.