Recover file fixes

[zonkmachine]

I'll collect a bunch of, hopefully, last tweaks to the recover file system here.

• Remove extra calls to autoSave() and a separate function runAutoSave() which was only related to the Limited Session that was dropped from the code in 290556e.

• Don't auto-save while playing by default. ( From criticism after RC3 where it created issues with glitches for users with weaker machines )

[tresf]

This PR had an unanswered question when it was merged.

[zonkmachine]

I missed that. Where? What was the question?

[tresf]

I missed that. Where? What was the question?

It's a commit review, so I can't hard-link it. Please scroll through the commit code, you can't miss it.

[zonkmachine]

Sorry, I force pushed and the comments seem gone. I can't find them in my mail either.

[tresf]

[zonkmachine]

When a review says Pending, it is only visible to you.
https://help.github.com/articles/reviewing-proposed-changes-in-a-pull-request/

Let's see, 7 days ago (review date) that really was a one line fix and the rest is removing old logic from the defunct Limited Session. I introduced void MainWindow::runAutoSave() to enforce saving whenever a new project was started to ensure there was a recovery.mmp. This would just be confusing to leave in as I've seen questions about why the recovery.mmp is spamming...

If you mean the last update from today before I rebased the PR, no. If you just change the bool "1" the update function still is hard coded "true" so I did it this way. Still hard coded but now false.

[tresf]

When a review says Pending, it is only visible to you.
https://help.github.com/articles/reviewing-proposed-changes-in-a-pull-request/

Thanks that explains it.

So when does auto-save kick in? Is it running in the background? What triggers it? Yes, the spamming of auto-save files is annoying and I tend to agree with that observation, but in my opinion, that's a failure to clean-up after successful save (or when explicitly click "Do not save").

If this feature causes performance issues, it's likely a threading prioritization issue.

[zonkmachine]

So when does auto-save kick in? Is it running in the background? What triggers it?

The timer is started here and the first save now is done after the first time-out of m_autoSaveTimer.

Two of the most common complaints around the auto-save function are both user related:

1. Error message: 'Could not open /recover.mmp for writing...'
   Most likely from the user setting the working directory to '/'.
2. User doesn't save at all but uses auto-save as backup and therefore doesn't have any real safety net so when the crash finally comes they hit the forum with a post like: 'I had the auto-save on and worked for 2 days without turning the computer off but then it crashed and all was lost... How do I get my project back?'