In the past I've always replaced spokes damaged like that fairly promptly but a few years ago it occurred to me that I hadn't seen any such spokes actually break. If they were going to break it would presumably be by initiating a fatigue crack which meant two things to me
a) it probably wouldn't happen immediately and
b) there ought to be ways of preventing the fatigue crack initiation
So on training wheels (where a spoke failure is not going to be critical if it does happen) I've tried another approach. What I've done is to remove the spoke if necessary and then dress the spoke damage with a swiss file and if necessary straighten/ peen the spoke surface with a hammer (which introduces compressive residual stresses into the spoke surface, not dissimilar to those left by the swaging process that is used during manufacture of DB spokes).
It looks dramatic if you file the round part of a 14G spoke into a 'D' shape but the thickness across the short side of the D needs to be a deal less than 1.25mm before you have reduced the cross section to the 64% which is comparable to that of the centre part of a 14/16G DB spoke . This means that even quite badly damaged spokes can be fettled if you want to do this. The spokes are then stress-relieved in the normal way.
You might ask 'why do this?' And the reason I suppose is twofold
1) it is an interesting experiment (and of some relevance if you are 'doing field repairs' as it were) and
2) in a workshop the correct spokes are not always readily to hand or even available.
On the latter point the spokes could be any length, 14G PG, 14-15G DB, 14-16G DB, 13-14G SB, 13G PG, 15G PG and could have different butt lengths and/or different elbow lengths, or be in silver or black finishes. There is more than one type of black finish too.
So far, using this approach 'no breakages' is an encouraging sign but in truth I have not put enough miles on such wheels to really conclude if it is or isn't a really sound repair or not.
In most wheels such damage is confined to a region below the spoke crossings so only half the DS spokes are likely to be damaged anyway. It is worth noting that using certain types of DB spoke on the DS means that the 16G part of the spoke may be damaged by the chain. For example Sapim DB spokes have very short butts. I have built wheels with Sapim on the NDS and DT (with longer butts) on the DS, figuring that the DT spokes might be slightly stiffer, but they are also less likely to be seriously damaged since a chain will most likely contact/mark the longer 14G section.
Maybe others could say what time/mileage intervals have passed between such damage occurring and a spoke breakage that has initiated at the damaged part of the spoke?
Previous (bitter) experiences with duff wheels tells me that if the wheels/spokes are bad then you can start to get fatigue failures between about 700 and 2000 miles or so. But these were not 'damaged spokes' per se, they may behave differently.
cheers