program category clean up

[rwscott@…]

Go one step farther to prevent mistakes such as invalid values being stored and used. The original patch, bff155970773, changed the category type from a string to an enum to help avoid mistakes in the future. It however meant that a couple of string arrays, and some static int's (added later to fix the fix), _must_ be kept in sync. This just changes the type of future mistakes. This defines the category type in one place only.

[rwscott@…]

Go one step farther to prevent mistakes such as invalid values being stored and used

[Karl Egly]

a variation on the patch

[Karl Egly]

Find attached my variation on the patch (as discussed back on ​http://irc.mythtv.org/ircLog/channel/4/2013-04-14:18:00 )

[stuartm]

I'm going to quote myself from IRC:

"stuartm: I can't really argue with the reasoning for the patch just sent to -dev, but I think such macros are as clear as mud :/"

...

"stuartm: I don't know how big a problem it really is, the category types haven't changed in years and no-one is proposing extending them any time soon, the aim of my commit was to avoid future bugs through typos or invalid assumptions such as the Dish EIT one – basically that you could stuff any old value into that field"

In summary, while I appreciate the patch, I think it hurts code readability. The content of that enum is never likely to change. The reason for switching to the enum in the first places was because the category types are used all over the code creating a substantial risk that invalid values might be used. The risk of someone adding to the enum without updating the array/int is insignificant by comparison.