1st draft: PRs fixed in 3.1.1

[Gabriel Dos Reis]

Joe Buck <Joe.Buck@synopsys.com> writes:

[...]

| > As you know, just because something used to "link" does not mean it
| > "worked".
| 
| Correct, but I personally used the feature to track down a memory leak
| problem in the past.  I am fully aware of the issues with the
| one-definition rule, and I would never ship a binary that mixed code with
| __USE_MALLOC with other code.  Just the same, once the change was made,
| Purify was immediately enlightening.  I understand why I now have to
| rebuild the library, but this was not required before.

I'm not convinced that "required" is the right term here.  I would
instead speak of "enforced".

I'm having trouble to qualify a hole that got fixed as a regression.  

-- Gaby