Fixed threading issue using a thread pool

[marco-calautti]

This addresses Issue #36
I tested it on Linux using GCC, and everything seems to be working fine. Needs testing on Windows with VC++.

[CookiePLMonster]

Looks good now, but needs squashing. I'll test on Windows later and report back.

[marco-calautti]

Just squashed the pull request.

[CookiePLMonster]

I just realized - since m_checksumJobs is being waited for when the file write/view "finishes", thread pool could probably be a global variable instead of instantiating pools over and over for each view.

[marco-calautti]

I just realized - since m_checksumJobs is being waited for when the file write/view "finishes", thread pool could probably be a global variable instead of instantiating pools over and over for each view.

Do you want me to do it?

[CookiePLMonster]

Do you want me to do it?

If you're up for it that'd be great. I haven't verified it locally but I see no reason why it wouldn't work.

[marco-calautti]

I made the thread pool global a the level of the IsoWriter. Maybe having it in the main is not nice.

[CookiePLMonster]

Yep, that's a better idea than making it global-global indeed. Conceptually it also makes sense - if multiple ISO writers ever coexist in the same process, it would make sense for their workers not to fight for the thread pool.

[CookiePLMonster]

Checked on Windows, everything works as expected.

[marco-calautti]

Great! I guess you are going to merge the pull request, then.

[CookiePLMonster]

I would if I could - I have no merge rights in this repository, only @Lameguy64 can merge.

[CookiePLMonster]

@marco-calautti For the time being I merged this PR to the develop branch of the fork we used to work on v2.00:
CookiePLMonster@3199947

I had to rebase your changes manually so git history might be imperfect - but you're welcome to open the same PR against the fork. In case you're interested, I created a new develop-old branch you could PR against, and then I'd erase my own merge and restore develop to clean state this way.

[marco-calautti]

I see the develop-old is some commits behind the master of this repo. Is that fine?

[CookiePLMonster]

Yes, the master of this repo contains some changes we've not "downstreamed" yet.

[marco-calautti]

@Lameguy64 any news in merging this?

[CookiePLMonster]

I don't think we're going to get an answer anytime soon, to be fair.