Boost format stray roundup

[johnhaddon]

This fixes a couple of strays that escaped the purge in #1514, and a couple of associated issues. I found these by removing support for boost::format() from the MessageHandler class. @ivanimanishi, if that is a step too far then just let us know and we can drop that commit from this PR. The single remaining boost::format() usage is now inFileSequence::fileNameTemplate(), which is a protected method. There are no derived classes that can make use of it in Cortex or Gaffer, so I wonder if IE has something internally using it, or if it should have been private all along. If the latter, then I'll remove it too.

[murraystevenson]

Thanks John! Changes look good from my perspective. Great to see we're almost free of boost::format()...

[johnhaddon]

@ivanimanishi, can I take your silence as approval? :)

[ivanimanishi]

We'll have to update our internal repositories, but if I understand correctly, if we start using fmt::format() in our code, it should still work with older cortex (10.5 and 10.6), right (since it would just be returning an std::string)?
If that's the case, then I'm ok with approving this PR.

[johnhaddon]

Yes, you'll be able to do msg( level, context, fmt::format( ... ) ) in all Cortex versions - as you say, it just returns a string. It is a reasonable effort to move all the boost::format() over to fmt::format() (I did most of it for Gaffer), but it's so much nicer once it's done. OK to merge then?

[johnhaddon]

I've pushed two more commits : one removes FileSequence::fileNameTemplate(), and the other fixes a bad format string in SceneCache. With this merged, we will be free of boost::format() everywhere.