So, that was an adventure.
First: I soldered one C2837, and one A1186. Powered up, Dim-bulb tester light bulb lit bright.
Second: I unsoldered the A1186, but left the one C2837 soldered, powered up, dim-bulb tester lit a little very briefly, then went off.
Third: I unsoldered the previously soldered C2837, and soldered the other C2837 and A1186. Powered up, and the dim-bulb tester lit bright. Unsoldered the A1186 leaving the one C2837 soldered, powered up, dim-bulb tester lit a little very briefly, then went off. Here's where things got exciting.
Fourth: I soldered the other C2837 so that I had the two C2837's, and no A1186. Powered up, the light lit briefly, then went off, and then the amp made the click sound like it does when it powers up, and sparks flew and the light lit bright. Couldn't shut down fast enough! Then I looked around until I found where the sparks came from. One C2837 looked like it shorted to the heat sink.
This is likely my fault. Quite a while back, I removed the transistors and cleaned the old dried up heat sink compound and remounted the transistors. I'm guessing they are supposed to be insulated from the heat sink? A quick google search revealed insulation pads for just this situation. I'm guessing that should be my next step, before anything else is done.
Interestingly enough, the transistor still tests good with my DMM.