If someone could verify this maybe with noogie? I can add files to /system/ but any modifications to an existing /system/ file causes a boot loop from what I can tell.
fwiw, I've also seen this behavior on 1.0.0 and 1.0.1 when replacing android system files on a fully running system. (dalvik cache?)
Just did a quick test on editing build.prop with noogie and then reboot.
Code:
#cat build.prop
...
ro.config.sync=yes
dalvik.vm.stack-trace-file=/data/anr/traces.txt
#Added by ros87
ro.dummy.test=hello
#
No problems with that.
Then I tried copying that patched framework.jar (again trough noogie) and rebooted.
As expected there were loads of complaints about incorrect dependency signatures by dalvik which would force it to rebuild cache.
In the end it booted just fine and on subsequent boots there's no complaints.
Unit works just fine with the patched framework, but I cannot confirm the patch itself working as I haven't tried installing gapps yet.
This is more or less the same behavior I've seen before which suggests to me there's nothing new and sinister added to 1.1

---------- Post added at 12:39 PM ---------- Previous post was at 12:33 PM ----------
Update:
I quickly wrote a small app that displays AndroidID, Phone Type and Device ID
With the patched framework getDeviceID() now returns a 15 digit number instead of null, so this confirms it working as intended.
I'll try to set up market next.
---------- Post added at 12:45 PM ---------- Previous post was at 12:39 PM ----------
Update #2:
It's also worth noting that Device ID (IMEI) is regenerated on every call, which could cause trouble with certain apps.
This has been discussed before, just wanted to emphasize that this problem is still there with this patch.