From: Michael Snyder <msnyder@redhat.com>
Date: Wed, 12 Apr 2006 19:41:42 -0700
I get confused just thinking about all these variants.
If you have an argument for why one is more correct in
this context, I'm certainly open to it. I can't really
convince myself that the current behavior is correct,
since deprecated_read_register_gen calls regcache_cooked_read,
which may call regcache_raw_read, which may call
target_fetch_registers *before* capturing the current value.
Which it seems like would defeat the purpose of the whole
exercise...
Any of these arguments for or against apply to the Solaris thread code
which does use regcache_raw_read() in this same exact situation.