add try catch for old mfxInit call #132

[xelement]

Issue

We are an app with tens of millions of users，recently we found some crash report show that mfxInit throw unhandled exception which do not be catched on user device，just see dump file below.

Solution

just add try catch for mfxInit call

How Tested

try catch do not need test

dump:
3a0b8cfd-431a-4e62-8f13-6e3cc42156be.dmp

[mav-intel]

Thanks for your PR. A couple of requests:

1. Please sign-off on the PR as indicated in CONTRIBUTING.md
2. Please fix the code formatting to align with style guidance
3. Please fix the commit message to align with style guidance. I recommend:
   Catch unhandled exceptions when calling old init

You can check formatting yourself by running script/test lint

[xelement]

I'm sorry I didn't read the CONTRIBUTING.md，I've already modified it，issue#132

[mav-intel]

How Tested

try catch do not need test

@xelement Can you confirm that you no longer see the exception thrown in your local environment after this change?

[xelement]

How Tested

try catch do not need test

@xelement Can you confirm that you no longer see the exception thrown in your local environment after this change?

It was not my local environment，it was feedback from our users. maybe we can hold this pr for sometime. Our latest app version already includes this change, but it will take some time to see if it has any effect

[mav-intel]

Waiting for results from local testing before moving forward

[jonrecker]

It was not my local environment，it was feedback from our users. maybe we can hold this pr for sometime. Our latest app version already includes this change, but it will take some time to see if it has any effect

Thank you for the update. The attached dump file and the call stack in #132 do not include a reference to libvpl.dll, so it was not clear how this code is being called. Also the source code line numbers in the call stack seem to line up with the MediaSDK implementation of MFXInit() and MFXInitEx(), so perhaps that test was run using MediaSDK dispatcher (there is significant code in common). Any additional info about how this PR was validated would be helpful.