<Swing Dev> JDK-8189938: Proposed patch

[Semyon Sadetsky]

On 02/08/2018 01:19 PM, Matthias Bläsing wrote:

> Hi,
>
> Am Donnerstag, den 08.02.2018, 12:06 -0800 schrieb Semyon Sadetsky:
>> Yet another question. Do you know what might call CoInitialize in a
>> different mode? It looks like we use STA everywhere in JDK.
> I tried to find some documentation how the FullTrustProcessLauncher of
> an UWP application works, but while the MSDN full of documentation
> about UWP, the lifecycle of a native application seems not to be
> covered. But my suspicion is this: The launching of an APPX package
> involves the runtime system creating an environment for the application
> and that seems to rely heavily on COM.
It looks like there is an internal reason. Probably for UWP apps only 
MTA is allowed because of their life-cycle (the app may be suspended).
> So I assume, that COM is initialized outside the java core and the java
> process get this inherited. (But this is speculation!).
>
> I base this on the fact, that CoInitializeEx succeeds and multi
> threaded, but fails appartment threaded and the JDK code holds exactly
> three files calling CoInitialize (D2DPipelineManager.cpp,
> ShellFolder2.cpp, PLATFORM_API_WinOS_DirectSound.cpp).
>
>> On 02/08/2018 11:55 AM, Semyon Sadetsky wrote:
>>> I meant the case when MTA COM service is actually used, it might cause
>>> issues since the JDK code invokes an another API which assumes
>>> synchronization. At least performance may be affected.
>>>
>>> We don't do separate testing of JDK in MTA mode especially on the
>>> regular base. That is why I wouldn't consider MTA mode as supported.
>>>
>>> So, before push the change I would make sure of the regression test
>>> suite run in MTA mode hasn't brought any surprises.
> Aggreed and I'd like to help you, but in fact I was happy, that I
> managed to compile the JDK on windows and I have zero experience
> working on the JDK sources.
Prasanta,

since the bug is assigned to you can you test the suggested fix in MTA 
mode and sponsor it if there are no issues?

--Semyon
>
> As an observation: The COM initialization and usage looked pretty tight
> together in the sources and at least initialization happens only in the
> three files mentioned above. The DirectSound part is even limit to a
> small thread. The ShellFolder2 code looks to be the biggest direct
> user.
>
> Greetings,
> Matthias