Fwd: Re: RFR 8056084: Refactor Hashtable to allow implementations without rehashing support

[Mikael Gerdin]

Hi Jon,

On 08/27/2014 06:54 PM, Jon Masamitsu wrote:
> Mikael,
>
> Just a resend since I forgot the mailing list on my
> reply.
>
> Jon
>
>
> -------- Original Message --------
> Subject:     Re: RFR 8056084: Refactor Hashtable to allow
> implementations without rehashing support
> Date:     Tue, 26 Aug 2014 13:45:14 -0700
> From:     Jon Masamitsu <jon.masamitsu at oracle.com>
> To:     Mikael Gerdin <mikael.gerdin at oracle.com>
>
>
>
> http://cr.openjdk.java.net/~mgerdin/8048268/hashtable-refactor/webrev/src/share/vm/utilities/hashtable.cpp.frames.html
>
>
> On right side
>
> 372 #endif // SOLARIS || CHECK_UNHANDLED_OOPS
>
> to
>
> 372 #endif CHECK_UNHANDLED_OOPS

As I wrote in my reply to Gerard I've decided to revert the #ifdef 
change since I can't really motivate doing it, it seems to be unnecessary.

Thanks
/Mikael

>
>
> Otherwise, looks good.
>
> Reviewed.
>
> Jon
>
>
> On 08/26/2014 08:26 AM, Mikael Gerdin wrote:
>> Hi all,
>>
>> For JDK-8048268 I need to do some refactoring of the Hashtable class
>> template.
>>
>>  From the bug description:
>> In order to simplify usage of the Hashtable<T, F> class the code to
>> perform
>> rehashing of the SymbolTable and StringTable should be factored out to an
>> intermediate class common to those two tables.
>> Also factor out HashtableEntry freelist allocation to allow
>> implementors an
>> easier route for providing their own memory management.
>>
>> Introducing the new intermediary superclass requires some calls to
>> superclass
>> functions to be qualified using the "this->" syntax due to the C++
>> template
>> name lookup rules.
>>
>> a comment about hashtable.cpp RehashableHashtable::dump_table:
>>
>> -  int bucket_bytes = (int)num_buckets * sizeof(bucket(0));
>> +  int bucket_bytes = (int)num_buckets * sizeof(HashtableBucket<F>);
>>
>> I decided to change the sizeof to count the size of the hashtable
>> bucket class
>> instead of the HashtableEntry* returned by bucket(0), even though they
>> may be
>> of the same size currently.
>>
>> Note that the webrev url refers to another bug.
>> Webrev:
>> http://cr.openjdk.java.net/~mgerdin/8048268/hashtable-refactor/webrev/
>> Buglink: https://bugs.openjdk.java.net/browse/JDK-8056084
>>
>> Testing: JPRT.
>>
>
>
>