concurrency saving problems (malformed JSON)

[smysnk]

Hopefully I am doing this correctly, created a new branch without my other modifications. This branch should solve the concurrency saving problems. I realize most people won't run into this issue but just incase!

[smysnk]

Still working on my concurrency skills, what you're suggesting sounds cleaner.

[tntim96]

Thanks.

Due to the multiple commits I've added this in manually, so let me know if I missed anything.

While not the ideal approach (saving data concurrently to separate files would be better) it is definitely an improvement, but may lead to web-request time-outs. I've committed to trunk but haven't created a separate release. Let me know if that is preferable.

[smysnk]

I am currently using a fork so no need to worry about a new release. Thanks

The saving/merging process would rarely take longer than a second, I don't think timeouts will be a problem.

[tntim96]

Actually, you might as well stick with the save/merge process for the time being. The above code won't work as the objects are created for each thread and need something else to use as a lock. In addition, even if it did work (by locking on JSONDataSaver.class for example) it will cause blocking even if the threads are saving to different report directories. I'll try to add a solution that only blocks when the report directory is the same.

[tntim96]

I think I've fixed this in trunk - i.e. it only blocks the thread if the report directory is already in use. My testing looks OK, but if you get a chance, let me know how it goes.