Re: RFR 8147491: module graph consistency checks after jlink plugins operate on module pool

[Sundararajan Athijegannathan]

Hi Mandy, Paul,

Thanks for the reviews. Updated webrev as per suggestions:

http://cr.openjdk.java.net/~sundar/8147491/webrev.02/

Thanks,
-Sundar


On 8/25/2016 10:50 PM, Paul Sandoz wrote:
>> On 25 Aug 2016, at 10:15, Paul Sandoz  wrote:
>>
>>
>>> On 25 Aug 2016, at 09:33, Mandy Chung  wrote:
>>>
>>>
 On Aug 25, 2016, at 7:18 AM, Sundararajan Athijegannathan 
  wrote:

 Please review http://cr.openjdk.java.net/~sundar/8147491/webrev.01/ for
 https://bugs.openjdk.java.net/browse/JDK-8147491
>>>
>>> I think ResourcePoolConfiguration::validate should simply use all modules 
>>> in the resource pool as the roots.  ImagePluginStack constructors don’t 
>>> need to know anything about the input root modules.
>>>
>> I am under of the mechanics of this aspect.
> s/under/unsure.
>
> (I need to dial down the aggressiveness of auto correction, some day it’s 
> gonna produce something far more embarrassing.)
>
> Paul.
>
>> It would nice to avoid the propagation of the Set but presumably the 
>> validation should not be using the roots after all plugins have been run, 
>> since a plugin could add a new module?
>>
>>
>>> Other than that, it looks fine.
>>>
>> +1
>>
>> Paul.

Re: RFR 8147491: module graph consistency checks after jlink plugins operate on module pool

[Mandy Chung]

> On Aug 25, 2016, at 10:15 AM, Paul Sandoz  wrote:
> 
> 
>> On 25 Aug 2016, at 09:33, Mandy Chung  wrote:
>> 
>> 
>>> On Aug 25, 2016, at 7:18 AM, Sundararajan Athijegannathan 
>>>  wrote:
>>> 
>>> Please review http://cr.openjdk.java.net/~sundar/8147491/webrev.01/ for
>>> https://bugs.openjdk.java.net/browse/JDK-8147491
>> 
>> 
>> I think ResourcePoolConfiguration::validate should simply use all modules in 
>> the resource pool as the roots.  ImagePluginStack constructors don’t need to 
>> know anything about the input root modules.
>> 
> 
> I am under of the mechanics of this aspect. It would nice to avoid the 
> propagation of the Set but presumably the validation should not be using the 
> roots after all plugins have been run, since a plugin could add a new module?

I think we are essentially saying the same thing.  All modules in ResourcePool 
should be validated as they are the set going to be linked in the image.  
Configuration::requiresRequiresAndUses with an empty set will resolve no 
modules.

Mandy

Re: RFR 8147491: module graph consistency checks after jlink plugins operate on module pool

[Paul Sandoz]

> On 25 Aug 2016, at 10:15, Paul Sandoz  wrote:
> 
> 
>> On 25 Aug 2016, at 09:33, Mandy Chung  wrote:
>> 
>> 
>>> On Aug 25, 2016, at 7:18 AM, Sundararajan Athijegannathan 
>>>  wrote:
>>> 
>>> Please review http://cr.openjdk.java.net/~sundar/8147491/webrev.01/ for
>>> https://bugs.openjdk.java.net/browse/JDK-8147491
>> 
>> 
>> I think ResourcePoolConfiguration::validate should simply use all modules in 
>> the resource pool as the roots.  ImagePluginStack constructors don’t need to 
>> know anything about the input root modules.
>> 
> 
> I am under of the mechanics of this aspect.

s/under/unsure.

(I need to dial down the aggressiveness of auto correction, some day it’s gonna 
produce something far more embarrassing.)

Paul.

> It would nice to avoid the propagation of the Set but presumably the 
> validation should not be using the roots after all plugins have been run, 
> since a plugin could add a new module?
> 
> 
>> Other than that, it looks fine.
>> 
> 
> +1
> 
> Paul.

Re: RFR 8147491: module graph consistency checks after jlink plugins operate on module pool

[Paul Sandoz]

> On 25 Aug 2016, at 09:33, Mandy Chung  wrote:
> 
> 
>> On Aug 25, 2016, at 7:18 AM, Sundararajan Athijegannathan 
>>  wrote:
>> 
>> Please review http://cr.openjdk.java.net/~sundar/8147491/webrev.01/ for
>> https://bugs.openjdk.java.net/browse/JDK-8147491
> 
> 
> I think ResourcePoolConfiguration::validate should simply use all modules in 
> the resource pool as the roots.  ImagePluginStack constructors don’t need to 
> know anything about the input root modules.
> 

I am under of the mechanics of this aspect. It would nice to avoid the 
propagation of the Set but presumably the validation should not be using the 
roots after all plugins have been run, since a plugin could add a new module?


> Other than that, it looks fine.
> 

+1

Paul.

Re: RFR 8147491: module graph consistency checks after jlink plugins operate on module pool

[Mandy Chung]

> On Aug 25, 2016, at 7:18 AM, Sundararajan Athijegannathan 
>  wrote:
> 
> Please review http://cr.openjdk.java.net/~sundar/8147491/webrev.01/ for
> https://bugs.openjdk.java.net/browse/JDK-8147491


I think ResourcePoolConfiguration::validate should simply use all modules in 
the resource pool as the roots.  ImagePluginStack constructors don’t need to 
know anything about the input root modules.

Other than that, it looks fine.

Mandy