General rules when fianchetto bishop is a good idea

I have been studiyng basic chess opening rules, develop toward the center, don't move same piece twice, don't bring the queen to early, castle as soon as you can. I read articles on various openings. Pro cons, whites ideas, black ideas.

There is one think that still escapes me. How to determine if fianchetto-ing the bishop is a good idea. Especially the bishop on whose side I will castle. I am aware that moving pawn in front of my castled king is weakening his defence and that only moving my rook pawn to prevent my opponent from parking his knight or bishop in front of my king and give my king escape route in case of back rank check is advisable.

In order to fianchetto if I am white for instance I need to play g3 which can be seen as weakening my kings position with out giving him escape route.

I am aware that some openings deliberately play this system and that it is considered good if fianchetto-ed bishop can be kept on the board.

But I still don't have any idea how can I determine in a given position is fianchetto a good idea. Are there any general rules I can use to evaluate the position and decide should I fianchetto bishop or not?

Thanks in advance

some general rules: when you fianchetto your bishop, don't block it with your own pawns, so if you play b3 and put your bishop on b2, then don't play d4 unless you can trade off that d pawn or advance it. Also, try not to leave a hole by playing for example e6 and g6. Instead you should try playing something like d6 and g6 because that doesn't leave a hole on f6, and it lets both bishops out.

I've never though about this, it's an interesting question.

The answer that comes to mind unfortunately is probably not very helpful.  The times it strikes me as a good thing to do (when it's not part of memorized moves) is when it fits in with an overall deployment of my forces.

As white, it's usually because I want added control of e4, d5, or c6, and fits in with some kind of idea in the center and or queenside.  But that it fits in well with the positioning of the rest of my pieces is important too.  If my other pieces can't affect the queenside and center, then just having the bishop there isn't useful.  Maybe in that case I'd look to trade it off instead of taking the time to play g3 and get by bishop to g2.

When the diagonal is clear of pawns, it's a good indication that it's a very good diagonal to have... the exception being if there's nothing going on on the queenside.

When pawns are locked for white on the light squares, e4 and d5 let's say, the finachetto isn't ideal, but it's not bad either.  I'd say solid but a bit passive.  If you're stopping your opponent from breaking in the center so that you can do something somewhere else... then it's definitely worth it.  But I might say you'd have to be careful not to trade off too many pieces and get left with that bishop by itself in the endgame.

Caution must be taken when commiting to a fianchetto early in a game on the flank you intend to castle, and especially if your opponent has the option to castle on the opposite side. In open positions this may allow your opponent develop his pieces in such a way as to use the fianchetto structure as an "anchor" for an attack e.g. Be3,Qd2, g4, h4-h5 in Yugoslav attack. 

However, there are excpetions such as KID attack where the fianchetto in combination to the locked centre and kingside spatial advantage allow black to develop a kingside attack without endagering his king.