Follow jsc in throwing an exception when using test or exec on a...

Follow jsc in throwing an exception when using test or exec on a
regexp with no input.

Fixed problem with assertThrows.

Deleted test that tests arbitrary limits on the sizes of regular
expressions.

Committed: http://code.google.com/p/v8/source/detail?r=907

[Christian Plesner Hansen, 2008-12-03 12:15:59]

Lgtm.

http://codereview.chromium.org/13088/diff/1/5
File src/regexp-delay.js (right):

http://codereview.chromium.org/13088/diff/1/5#newcode318
Line 318: function RegExpGetInput() { return regExpInput || ""; }
Note that this changes the behavior when input is a false-like value like false,
0 or null.  Consider using IS_UNDEFINED instead.

Also, do we want to distinguish between regExpInput being uninitialized and
regExpInput being initialized and containing undefined?

http://codereview.chromium.org/13088/diff/1/3
File test/mjsunit/mjsunit.js (right):

http://codereview.chromium.org/13088/diff/1/3#newcode93
Line 93: var threw_exception = true;
I believe we usually use lower camelcase in JS, lower case with underscores is
for C++.

[Mads Ager (chromium), 2008-12-03 12:43:07]

On 2008/12/03 12:15:59, Christian Plesner Hansen wrote:
> Line 318: function RegExpGetInput() { return regExpInput || ""; }
> Note that this changes the behavior when input is a false-like value like
false,
> 0 or null.  Consider using IS_UNDEFINED instead.

Done.  Using IS_UNDEFINED is clearer.
 
> Also, do we want to distinguish between regExpInput being uninitialized and
> regExpInput being initialized and containing undefined?

We do not have to because the setter always converts the argument to a string. 
The comment in the code is wrongly stating that we do not convert to string on
assignment and I have fixed that.

> Line 93: var threw_exception = true;
> I believe we usually use lower camelcase in JS, lower case with underscores is
> for C++.

Fixed.