I have a "two-part" answer to your question:
1) The value of Tg is the point where the material begins to change more radically, but that doesn't mean it is not acceptable to exceed that temperature. For example, if you measured the Z-axis expansion of a material as you raise the temperature, you would see a graph that is roughly linear up until it approaches the Tg temperature, and then after that the curve (or ramp) would change more radically. Just because the material is expanding at a different rate aboove Tg doesn't mean you can't exceed that temperature for short periods of time to accomplish your soldering. Some articles you read will discuss the fact that after several thermal excursions that material will start to degrade (well, technically after each thermal excursion the material has degraded), which is why the number of rework cycles is often held to a limit
2) Although you say "normal" FR4 is Tg 125, I think if you polled a few fabricators you would find that people are "typically" using a higher grade. We use materials that are more in the range of 150-170, and lead free designs are calling for Tg 170 and up.
p.s. You probably already know this, but you should never design a product to run at a normal operating temperature near the Tg for your material. For example I think a 170 Tg material should never be operated above (something around) 135 for any length of time.
hope that helps, Jack
Thanks Jack. I got better understanding of Tg temperature from your answer.
I realize I didn't answer one of your main questions about components, but that is not an area that I know much about.
There is another forum hosted by the IPC called TechNet (not trying to drive people away from here, but it has over 1200 members already)
Standard FR-4 should not be used in this case, as you point out the Tg is way too low.
You do need to use a substrate with a higher Tg like 370 HR. I use this and FR406 frequently
on RoHS designs.