A few remarks:
In the 80s Eskom had its own coal mines, so it could manage its own supply. Apparently this isn't so any more. Maybe not a very smart decision to cut off one's own supply chain...

I think the main issue in South Africas electricity supply are the extremely low consumer prices. As far as I remember they were at 1 cent per kWh before they doubled to 2 ct/kWh (compare: average in Germany: 20 ct/kWh or 30 US-ct/kWh). Thus, energy providers are far from being motivated to invest in more capacity, which would produce anything but losses for them. It is just like in the socialist countries where people had sort of a flat rate on energy, so nobody cared about energy efficiency. As for South Africa I remember some people even left their air conditioning on when they were on holidays.
Of course it will be very hard for the government to wean the people from the good old cheap rates - people are already now protesting against rising petrol prices. But sooner or later there won't be any choice as the resources as such become more expensive to dig up.

I have the impression that South African coal may peak quite soon:
There were very large deposits exploitable from open pit mines, but the shallow ones are almost dug up, and the deeper ones have to be extracted from underground mines. This is especially difficult with the geological conditions in SA, where the occurrence basalt dykes may obstruct the production process. Here, underground coal gasification sounds like a good solution. But as far as I know this technology is still in an experimental stage, and all former projects trying to use it were discontinued. So if UGC will ever work it will probably take quite a while until it will bring a considerable contribution to energy supply, and this will be probably post-peak.
Furthermore I don't see that UGC will have a better carbon balance - and adding the cost of CSS it will be very doubtful if coal power plants will be profitable at all - certainly not at a price of 2 cent per kWh!

Could you expand on the difficult geologic conditions? The basalt dikes you mention, do these cross cut the coal seams?

Yes, they do. If you are lucky you can localize them with a magnetometer (20 years ago I did this walking through the bushveld, now this is also done with helicopters). If you miss one you may be out for a bad surprise. Whereas in an opencast mine you just blast the whole stuff and pick the coal out, in an underground mine a dyke may provide a much darker outlook: A 2007 study for the EU "COAL OF THE FUTURE" has a few more details (currently offline):
"In southern African mines the coal seams are heavily intruded by Jurassic dolerites dykes and sills. These either burn the coal or cut through it and due to their hardness and possible seam displacements, which makes mine planning difficult.
The presence of these intrusives is the main reason why so few long wall or short wall mining methods are used in South Africa."

This makes mine planning even more complex than they are already: Unlike the regular (although folded) coal seams we are used in Europe the South African seams I have seen were much more irregular due to a different deposition history. It was quite hard for me to keep the track and I was impressed of the local geology team who managed it regularly.