Downport JDK-8212828 (posix_spawn on Linux as non-default) to 11u
David Holmes
david.holmes at oracle.com
Fri Nov 30 03:34:38 UTC 2018
Hi Joe,
On 30/11/2018 12:14 pm, Joseph D. Darcy wrote:
> Hello,
>
> The specification of a system property includes both its name and its
> set of legal values. If the set of legal values changes, a CSR is
> warranted to document that.
Hmmm. That should be clearly documented on the CSR page(s). At the
moment the only text is:
- Changing or defining a new system or security property
There's no clear notion of a "set of legal values" in general because it
will depend on the property and how the value of the property is used.
And in this case we're not adding a new value IIUC just allowing an
existing value to also be used on Linux.
The allowed values of the property, per platform, will need to be
documented separately regardless of a CSR request.
A CSR request for this seems excessive overhead to me. But Joe is the
CSR lead so ...
Cheers,
David
> Cheers,
>
> -Joe
>
> On 11/29/2018 4:26 PM, David Holmes wrote:
>> Hi Roger,
>>
>> On 30/11/2018 5:54 am, Roger Riggs wrote:
>>> Hi Thomas,
>>>
>>> We'll also need a retroactive CSR for JDK 12 to change its status in
>>> JDK 12 to be documented.
>>
>> With my CSR Group member hat on ... not sure what you mean by
>> "documented" here? The property already exists and should already have
>> had a CSR (or old CCC) request when it was added. I don't think
>> allowing a new value for that property requires a CSR request, simply
>> an update to the documentation of the property. Of course if that
>> documentation is part of an API then that itself would need a CSR
>> request - but I don't think we document this at the API level.
>>
>> David
>> -----
>>
>>> Can you take a swing at that?
>>> On the Jira issue[2] there is a "More" menu with a 'Create CSR' button.
>>> It comes pre-filled with a template that should be pretty
>>> straightforward to fill out.
>>>
>>> A CSR will also be needed for JDK 11 (pretty much a copy).
>>> In the compatibility section, point out that it does not change the
>>> default.
>>>
>>> I created a draft release note for JDK 12 [1].
>>> Please review and comment or edit it directly with suggestions.
>>> There is a release note style guide for reference [3].
>>>
>>> Thanks, Roger
>>>
>>> [1] https://bugs.openjdk.java.net/browse/JDK-8214454
>>> [2] https://bugs.openjdk.java.net/browse/JDK-8212828
>>> [3[
>>> https://wiki.se.oracle.com/display/JPG/OpenJDK+Release+Note+Style+Guide
>>>
>>> On 11/28/2018 11:56 AM, Thomas Stüfe wrote:
>>>> On Wed, Nov 28, 2018 at 5:39 PM Alan Bateman
>>>> <Alan.Bateman at oracle.com> wrote:
>>>>> On 28/11/2018 15:36, Thomas Stüfe wrote:
>>>>>> Hi all,
>>>>>>
>>>>>> I would like to have this patch in 11u too, so I just did a 11u
>>>>>> downport request.
>>>>>>
>>>>>> Our agreement was that launchMechanism=POSIX_SPAWN would stay as
>>>>>> optional non-default value for jdk12, and at the start of JDK13
>>>>>> development we would switch the default to posix_spawn.
>>>>>>
>>>>>> I would like to see this non-default-but-possible-to-set option in
>>>>>> 11u
>>>>>> too, just to broaden the test surface. We'd get more testers that
>>>>>> way.
>>>>>>
>>>>>> What do you think, does that make sense?
>>>>>>
>>>>> It would be nice to have gotten feedback from people trying it out
>>>>> with
>>>>> JDK 12 EA builds but I don't know if that is possible.
>>>>>
>>>>> If it is backported then a CSR will be required as it will be adding a
>>>>> new supported value. A release note or other documentation updates
>>>>> might
>>>>> be needed too.
>>>>>
>>>> Oh, you are right. Do we have a release note/documentation for jdk12?
>>>>
>>>> ..Thomas
>>>>
>>>>> -Alan
>>>
>
More information about the core-libs-dev
mailing list