RFR (XXS): 8006105: [parfait] Possible null pointer dereference at hotspot/src/share/vm/c1/c1_LIRGenerator.cpp

[Morris Meyer]

Albert,

In many of the other parfait issues I closed recently - I used 
guarantee() to assert pointers.

         --mm

On 4/29/13 12:39 PM, Albert Noll wrote:
> Hi Roland,
> thank you very much for the feedback.
> How about using guarantee(...) to ensure that src_con != NULL?
>
> Best,
> Albert
>
> ----- Original Message -----
> From: roland.westrelin at oracle.com
> To: albert.noll at oracle.com
> Cc: hotspot-compiler-dev at openjdk.java.net
> Sent: Monday, April 29, 2013 5:43:09 PM GMT +01:00 Amsterdam / Berlin / Bern / Rome / Stockholm / Vienna
> Subject: Re: RFR (XXS): 8006105: [parfait] Possible null pointer dereference at hotspot/src/share/vm/c1/c1_LIRGenerator.cpp
>
> Hi Albert,
>
> It doesn't look like a correct fix to me. src_con == NULL shouldn't be possible here AFAICT. It would mean that src is not of type object which shouldn't happen so by making this change we're actually allowing a broken return value to go unnoticed.
> This said, I'm not sure what code change would prevent parfait from complaining while keeping the code correct.
>
> Roland.