/[cmucl]/src/general-info/release-19d.txt
ViewVC logotype

Diff of /src/general-info/release-19d.txt

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.24.2.1 by rtoy, Mon Oct 2 18:38:54 2006 UTC revision 1.24.2.2 by rtoy, Fri Oct 6 14:44:01 2006 UTC
# Line 59  New in this release: Line 59  New in this release:
59          number with >= 106 bits of precision (about 33 digits).  Known          number with >= 106 bits of precision (about 33 digits).  Known
60          issues:          issues:
61    
62          * If you are expecting IEEE-style behavior, you don't get it:          * IEEE-style floating point behavior is not supported:
63            - signed zeroes aren't really available.            - signed zeroes exist but arithmetic operations may not
64                produce the correct signed zeroes.
65            - overflows don't return infinity but return NaN instead.            - overflows don't return infinity but return NaN instead.
66            - rounding might not be quite the same as IEEE            - rounding might not be quite the same as IEEE.
67            - SQRT is not accurate to the last bit, as required by IEEE.            - SQRT is not accurate to the last bit, as required by IEEE.
68          * Multiplying by a number very close to          * Multiplying by a number very close to
69            most-positive-double-float will produce an error even if the            most-positive-double-float will produce an error even if the
70            result does not overflow.  (This is an artifact of how            result does not overflow.  (This is an artifact of how
71            multiplication is done.  I don't have a solution to this.)            multiplication is done.)
72          * Read/write consistency is not working.  (Because conversion          * Read/write consistency is not working.  (Because conversion
73            from a bignum to a double-double-float doesn't really            from a bignum to a double-double-float doesn't really
74            understand the internal double-double-float format.)            understand the internal double-double-float format.)
75          * INTEGER-DECODE-FLOAT and SCALE-FLOAT aren't "inverses".          * INTEGER-DECODE-FLOAT and SCALE-FLOAT aren't "inverses".
76            That is, you can't take the result of integer-decode-float            That is, applying the result of integer-decode-float to
77            and use scale-float to produce exactly the same number.            scale-float does not produce exactly the same number.  This
78            This is because of how bignums are converted to            is because of how bignums are converted to double-doubles.
           double-doubles.  
79          * FLOAT-DIGITS always returns 106 even though there could be          * FLOAT-DIGITS always returns 106 even though there could be
80            more bits.  (Consider the double-double (1d0,1d-200)).  This            more bits.  (Consider the double-double (1d0,1d-200)).  This
81            will show up in PRINT where the printed result will have way            will show up in PRINT where the printed result will have way
# Line 83  New in this release: Line 83  New in this release:
83            number back won't give the same value.            number back won't give the same value.
84          * There is probably more consing than is necessary in many of          * There is probably more consing than is necessary in many of
85            the standard Common Lisp functions like floor, ffloor, etc.            the standard Common Lisp functions like floor, ffloor, etc.
86          * The special functions are not fully tested.  I did a few          * The special functions are not fully tested.  A few random
87            random spot checks for each function and compared the            spot checks were done for each function and maxima was used
88            results with maxima to verify them.            to verify the results.
89          * The branch cuts for the special functions very likely will          * The branch cuts for the special functions very likely will
90            not match the double-float versions, mostly because we don't            not match the double-float versions, mostly because we don't
91            have working signed zeroes.            have working signed zeroes.
92          * Type derivation for double-double-floats might not be          * Type derivation for double-double-floats might not be
93            working quite right.            working quite right.
94          * PI is still a double-float.  If you want a double-double          * PI is still a double-float.  The double-double version of pi
95            version of pi, it's KERNEL:DD-PI.  (Soon to be EXT:DD-PI.)            is EXT:DD-PI.
96          * There are probably still many bugs where double-double-float          * There may be bugs where double-double-float support was
97            support was overlooked.            overlooked.
98          * The double-double arithmetic operations can be inlined by          * The double-double arithmetic operations can be inlined by
99            specifying (SPACE 0).  Otherwise, they are not inlined.            specifying (SPACE 0).  Otherwise, they are not inlined.
100            (Each double-double operation is about 20 FP instructions.)            (Each double-double operation is about 20 FP instructions.)
# Line 131  New in this release: Line 131  New in this release:
131        - An error is signaled if a declaration is used as the name of a        - An error is signaled if a declaration is used as the name of a
132          deftype, condition, or defstruct, and vice versa.          deftype, condition, or defstruct, and vice versa.
133        - An error is signaled when trying to generate a namestring from        - An error is signaled when trying to generate a namestring from
134          a pathname with just a version component (other than nil,          a pathname with just a version component (other than NIL,
135          :newest, or :unspecific).  CMUCL cannot print that readably.          :NEWEST, or :UNSPECIFIC).  CMUCL cannot print that readably.
136        - FLET and LABELS functions will catch errors in keyword        - FLET and LABELS functions will catch errors in keyword
137          parameters.  Previously, a keyword of NIL was silently          parameters.  Previously, a keyword of NIL was silently
138          accepted.          accepted.
# Line 173  New in this release: Line 173  New in this release:
173          out to C varargs functions.  Now we always copy any float args          out to C varargs functions.  Now we always copy any float args
174          to the corresponding int regs (or stack) as required by the          to the corresponding int regs (or stack) as required by the
175          ABI.  This isn't necessary for non-varargs functions, but          ABI.  This isn't necessary for non-varargs functions, but
176          CMUCL doesn't know functions which are varargs functions.          CMUCL doesn't know which functions are varargs functions.
177        - Callbacks with long-long args or results should work correctly        - Callbacks with long-long args or results should work correctly
178          now for Darwin/ppc.          now for Darwin/ppc.
179        - DESCRIBE no longer depends on having PCL loaded.        - DESCRIBE no longer depends on having PCL loaded.
180        - Tracing with no encapsulation appears to be working now for        - Tracing without encapsulation appears to be working now for
181          ppc.          ppc.
182        - A simple interface to sysinfo(2) has been added for sparc.        - A simple interface to sysinfo(2) has been added for sparc.
183          This is used to provide better values for MACHINE-TYPE and          This is used to provide better values for MACHINE-TYPE and
# Line 212  New in this release: Line 212  New in this release:
212        - (expt 1 <big number>) doesn't trigger a continuable error        - (expt 1 <big number>) doesn't trigger a continuable error
213          anymore and returns 1 immediately.          anymore and returns 1 immediately.
214        - Disassembling methods doesn't produce a type error anymore.        - Disassembling methods doesn't produce a type error anymore.
215        - The unknown condition type 'LISP:SOCKET-ERROR has been fixed.        - The unknown condition type LISP:SOCKET-ERROR has been fixed.
216          It properly signals the EXT:SOCKET-ERROR condition now.          It properly signals the EXT:SOCKET-ERROR condition now.
217        - The accuracy of the trig functions (sin, cos, tan) for large        - The accuracy of the trig functions (sin, cos, tan) for large
218          arguments has been improved for x86 and ppc.  Sparc already          arguments has been improved for x86 and ppc.  Sparc already
# Line 226  New in this release: Line 226  New in this release:
226          been fixed.  (CMUCL would sometimes crash to ldb about weird,          been fixed.  (CMUCL would sometimes crash to ldb about weird,
227          invalid objects.)  There are, however, still issues with weak          invalid objects.)  There are, however, still issues with weak
228          pointers.          pointers.
       - Hash table entries with a key and value of :EMTPY now work as  
         expected.  
       - EXT:READ-VECTOR can read binary data from streams with element  
         type BASE-CHAR or CHARACTER.  
229    
230    * Trac Tickets    * Trac Tickets
231        3. without-package-locks doesn't work with defmacro        3. without-package-locks doesn't work with defmacro
# Line 264  New in this release: Line 260  New in this release:
260          the value of h_errno is returned.          the value of h_errno is returned.
261        - A warning is printed when creating a weak key hash table with        - A warning is printed when creating a weak key hash table with
262          a test different from EQ.          a test different from EQ.
263          - An SCM and project management system is now available at
264            http://trac.common-lisp.net/cmucl.  Please use this to submit
265            tickets, documentation, project requests, etc.
266    
267    * Improvements to the PCL implementation of CLOS:    * Improvements to the PCL implementation of CLOS:
268    
269    * Changes to rebuilding procedure:    * Changes to rebuilding procedure:

Legend:
Removed from v.1.24.2.1  
changed lines
  Added in v.1.24.2.2

  ViewVC Help
Powered by ViewVC 1.1.5