The "new" RoHS (Restriction of Hazardous Substances)compliant solders (Certainly in the E.U., not sure about the United States or Canada) are effectively nearly pure tin with about 5% of various other metals like silver, copper, antimony etc. added to prevent it turning into alpha tin
and crumbling away.
Pure tin only has a melting point of 232*C, (450*F). Addition of small amounts of other metals will change this, but if the additions are less than 5% by mass, the change won't be much.
Pure lead on the other hand had a melting point of 327*C, (621*F). It is often alloyed with a small amount of antimony which makes it harder but does not alter the melting point much. In this base type, the blob of metal that locks the cap in place and acts as the electrical attachment of the lead-in wire looks
like lead and behaves like lead when melted, thus I assume it is either pure lead, (by the way the side cutters cut through it) or an alloy with very high lead content and a melting point near to that of lure lead.....which is generally high enough to withstand the cap/base temperatures of most lamps which generally don't exceed 250*C, (482*F).
The new tin based solders possess melting points closer to that of pure tin, the melting point of which is generally less than or near to the cap up temperatures of many traditional lamps.
So the strength of the glass is not an issue. It is sort of like saying that the Government has decreed that you MUST use hexagonal wheels on your car, so will changing the engine from a petrol, (gasoline) one to a Diesel one remove the issue of the, now bumpier ride?
One interesting side point to note here....the new RoHS solders...now used in all manner of consumer electronics, are harder and more crystalline/brittle than the traditional eutectic 60/40 Pb/Sn solder, (which stoicimetrically works out to be "Lead-IV Stannide") which had a malleability somewhat between that of pure lead and pure tin. The tendency for tin to alter itself to its non-metallic alpha form is still prevalent despite the additions of small quantities of other metals.
There are many "lead-free" these days solders but all contain at least 85% tin and the great proliferation of them seems to be as a result of manufacturers desperately trying to get around the alpha tin issue and a more reliable substance that won't crumble away like pure tin does...but also avoid infringing upon each other's patents. While this is happening one gets the sense that the problem initially created by Government Decree has not really been solved. These days we seem to be "re-creating" problems that our ancestors had already solved. 60/40 solder was a mature
technology, it had been tested to the "n-th degree" in both the lab and in the field and its long term statistical properties were well known and understood. But this new stuff has not attained that state yet ans we are part of the "testing" process everytime a piece of consumer electronics in our possession fails.
As a consequence, consumer electronic equipment is now far more prone to "dry" joints, fractured solder joints, often due to masgnetostriction and often found under P.C.B. mounted transformers, (which are heavy and vibrate) leading to lots of failure issues....HOWEVER, non-consumer electronics, D-Slams, Backhaul links, Servers, Transmitters and the like are exempt from the new RoHS decrees as the mean time between failures was deemed unacceptable in these more "important" applications if RoHS compliant solder was used...so they were exempt and still use traditional 60/40 solder of known and definable characteristics, particularly over time.
"Oh, you have to remove the potential to poison people or the environment , but not if in infringes on "national security" or "population control"