patches to move to gcc-3_2-branch

[Loren James Rittle]

>> libstdc++/8708 (__USE_MALLOC regression)

Upon looking at it closely, it is really hard for me to consider this
a regression.  User's use of __USE_MALLOC has always been problematic
in v3.  That report inspired an overhaul; I think it is a great new
debugging feature and removal of another problematic macro.  I
probably would have moved it without any thought but the merge of
include/bits/stl_alloc.h produced a large conflict because many
patches were made to formatting, etc on mainline.  It was upon
resolving that conflict that I reviewed it closely and concluded as
above.  FYI, in case it will save time later (i.e. someone else wants
this to be considered a regression), here are the transfer steps I
started to use:

cvs update -j1.24 -j1.25 docs/html/23_containers/howto.html
cvs update -j1.23 -j1.24 docs/html/ext/howto.html
cvs update -j1.11 -j1.12 include/backward/alloc.h
cvs update -j1.517 -j1.518 include/bits/c++config
cvs update -j1.24 -j1.25 include/bits/stl_alloc.h
(Resolve 1K+ line conflict.  However, it appears easier to redo by hand.)
cvs update -j1.14 -j1.15 src/stl-inst.cc
cvs update -j1.9 -j1.10 testsuite/21_strings/capacity.cc
cvs update -j1.1 -j1.2 testsuite/ext/allocators.cc
(Manually take a part of:)
cvs update -j1.21 -j1.22 config/linker-map.gnu