NOTE: Testing this could put your phone into a state where reboot doesn't work and you have to pull the battery. Some users reported symptoms like this before. DO NOT TEST THIS unless you are comfortable with that risk.
This is not a new bug - but reports about it were never very controlled they involved too many variables.
I have tested today and found that I can lock up the Luna UI by creating a reply chain of emails back and forth with another user. My test bed was a VZW Pre Plus running 188.8.131.52 and a FrankenPre2 (Sprint Com board in VZW Pre2) running 2.1. My friend with the Pre2 sent an email containing nothing more than the number 1; then I sent back the number 2; on a few messages we were more creative with things like "fore" for 4.
Eventually when I received message 15 and hit reply (making this the 8th back and forth) my phone became so unresponsive that it appeared locked up. The card containing the email could not be flicked away - screen taps would appear only after a l-o-n-g delay (a minute?). I was able to press OPT/SYM/R to reboot - however since I was also running F102A with OMG? VDEMAND I think something went wrong because it seemed to go into a fade to black death rather than a reboot. To be fair I had waited several minutes before I rebooted and the VDEMAND code may have been really ****ed off - I don't know.
I do know that opening the same email after a fresh reboot with VDEMAND set to MEDIUM still locked the UI up and when rebooted it did NOT cause the fade to black under voltage failure. On that reboot I reacted more quickly and didn't wait minutes.
I then tried sending a number of emails further using my PC to my friends Pre2 - it never crashed. There were some slowdowns but nothing significant.
However I can also confirm that this seems to have altered the email enough not to crash my phone. The emails sent afterwords using the PC browser when opened never crashed my phone. The 15th email when I reply still crashes it.
I have all the emails, and would be willing to work with someone further to validate these results.
Considering this has been around for at least a year I'm kinda sad that Palm never fixed it. I can guess that this is related to webkit and the improvements/changes in webkit in 2.1 seem to have resolved this.
One more item of interest. My mobile hotspot was in use during this test - the Pre2 was using my internet to do this test. The mobile hotspot did NOT stop working until I rebooted. This is what leads me to think this is LUNA not Linux that is hosed and since LUNA relies on webkit to render much of the OS UI this seems to make sense.
Additional items that could be tested:
Open a terminal window and monitor linux swap use.