When accounting for the octets left in the in-buffer that we haven't
read (or converted to characters), we were subtracting the index from
the total in-buffer length.  This is wrong if the file is less than
the total in-buffer length.  We should have subtracted from the actual
number of octets in the in-buffer.

changed.

* Rename the slot to %READTABLE-CASE (from READTABLE-CASE).
* Add READTABLE-CASE and (SETF READTABLE-CASE) functions, as required.
* Check for the standard readtable in (SETF READTABLE-CASE).

error.lisp::
* Create two new conditions, one for modifying the readtable and one
  for the pprint dispatch table.

exports.lisp::
* Export the two new conditions.

pprint.lisp::
* Add check to SET-PPRINT-DISPATCH to disallow modifying the standard
  pprint dispatch table.
* Allow PPRINT-INIT to modify the standard pprint dispatch table.

print.lisp::
* In WITH-STANDARD-IO-SYNTAX, don't copy a new dispatch table; bind
  *print-pprint-dispatch* to the standard table.

reader.lisp::
* Add check to disallow modifying the standard readtable.
* Allow INIT-STD-LISP-READTABLE to modify the standard readtable.

* Remove support for :athlon in fp-pop.
* Remove all of the enhanced pentium pro fp comparison code.  Use the
  non ppro everywhere.

one.

code/misc.lisp::
* Don't register :PENTIUM feature.

code/time.lisp::
* Replace :pentium with :x86; we assume all x86's have the rdtsc
  instruction.

compiler/x86/float.lisp::
compiler/x86/system.lisp::
* Remove the vop guard on :pentium; assume we're always running on a
  pentium or better.

code/misc.lisp::
* Remove :i486 from *features*

code/multi-proc.lisp::
* Change :i486 to :x86 on the off chance that multi-proc will be
  ported to other archs.

compiler/x86/cell.lisp::
compiler/x86/macros.lisp::
compiler/x86/memory.lisp::
* Remove guard on backend-featurep :i486 since we got rid of :i486 and
  have it always enabled now.

srctran.lisp::
 * The optional convert-type arg is always true, so remove it and
   update the local functions to convert always.
 * Clean up some comments as well.

float-tran.lisp::
 * Fix up one place where we were supplying a value for the optional
   convert-type arg.

accordingly.

 lispinit.lisp::
 * Remove :COMMON from *FEATURES*
 misc.lisp::
 * Remove unused defvar of *features*.

constants.

Without this, the general shift vop is used which has to test for the
sign of the shift before shifting.  This micro-optimization removes
the test when we know the sign of the shift.

 code/export.lisp::
 * Export new symbols

 compiler/x86/parms.lisp::
 * Define constants for useful the number of lowtag bits and masks and
   the number of fixnum tag bits and masks.

If the version is not specified, determine a default version based on
the git hash returned by git describe.  Thus, for tagged snapshot
builds, the version will automatically be chosen from the tag, with
"snapshot-" removed.

 src/contrib/asdf/asdf.lisp::
 * Update

 src/general-info/release-20e.txt::
 * Update

 code/float-trap.lisp::
  * In the default case, check FOP to see if it's defined.  This means
    it's a divide exception.

 code/x86-vm.lisp::
  * Also check to see if the offending instruction is a DIV or IDIV,
    which means we got an integer overflow.  Return the appropriate
    values in this case.

Apply the expt transform in more places.  The test script in the
ticket now passes.

The reader conditionals in {{{NOT-MORE-CONTAGIOUS}}} were incorrrectly
handling the case when the second type was a double-float. It was
always returning T in that case, but this is wrong if the first type
is a double-double-float.

This allows logical pathnames (which always have a host to parse
against) and search-lists (which don't have an explicit pathname to
parse against) to be parsed correctly.  See the comment for some
possible issues.