Testing changesets with OpenJDK6 bug ids

[Joe Darcy]

On 3/14/2013 10:01 AM, Andrew Hughes wrote:
> ----- Original Message -----
>> Hi,
>>
>> I have just pushed changesets [1] that will allow us to use bug ids
>> from
>> the OpenJDK6 JIRA [2] when pushing to the jdk6 forest.
>>
>> I would like to make sure that this is in fact, working. The attached
>> patch (sorry, webrev is not playing nice with me today) is a NOP
>> change
>> that adds and removes a file from jdk6/hotspot repo to verify that
>> the
>> commits are working.
>>
>> Andrew, can I list you as a reviewer?
>>
> I don't see any problem with the patches as a test and you can list me
> as reviewer too if you wish.
>
> Wouldn't it be better to use some leading zeros in the bug ID with a
> view to future formatting? i.e. 000001 rather than just 1?
>
>> Kelly suggested that _only_ bug ids from the OpenJDK6 JIRA should be
>> used once we switch over. Does this sounds sensible to others? Any
>> other
>> suggestions to avoid confusion about where to look for more
>> information
>> on a bug.
>>
> I don't see how this would work.  In fact, it'll already be broken by
> the patch I approved for Jiri.  Most of the patches are going to be
> backports with Oracle bug IDs.  I don't see many cases where you'd
> commit something ONLY to 6 and not 7 or 8.
>
> I wonder if it's worth having an identifier at the front to separate
> the bug IDs from the purely numerical Oracle ones?  So maybe OJ0001
> or OJ60001 though that looks a bit ugly...
>
>

If you want to stick with just numbers, there are numerical ranges that 
will not conflict with any Sun/Oracle bug such as 3xxxxxx. (If you have 
sufficient configuration control over your JIRA instance, you can set 
the starting bud id.) You could also use the full JIRA id of your bugs, 
$PROJNAME-NUMBER.

-Joe