GCC Status Report (2004-03-09)

[Richard Kenner]

    Before I start thinking about a replacement, I'd like to understand
    what I'll be trying the replace.  It appears that no-bo-dy can tell
    what is the purpose of RTX_UNCHANGING_P.

Well I can tell you the *purpose*, but if you want the *definition*, that's
something else entirely.

The original idea was to separate objects that we knew could not be
modified by stores, like constant pool objects.

Then, at some later point, it was realized that we could also optimize by
allowing a *single* store to "constant" object to initialize it, but things
started rapidly going downhill once we started to take more and more
advantage of that property.