;;;; -*- coding: utf-8; fill-column: 78 -*-
-changes relative to sbcl-1.0.56:
+changes relative to sbcl-1.1.8:
+ * enhancement: disassemble now annotates some previously missing static
+ functions, like LENGTH.
+ * optimization: calls to static functions on x86-64 use less instructions.
+ * optimization: compute encode-universal-time at compile time when possible.
+ * optimization: when referencing internal functions as #'x, don't go through
+ an indirect fdefn structure.
+ * optimization: SLEEP doesn't cons on non-immediate floats and on ratios.
+ * optimization: (mod fixnum) type-checks are performed using one unsigned
+ comparison, instead of two.
+ * optimization: enable more modular arithmetic transforms in the presence of
+ conditionals.
+ * optimization: bitwise OR forms can now trigger modular arithmetic as well,
+ when the result is known to be negative.
+ * optimization: recognize more cases of useless LOGAND/LOGIOR with constants.
+ * optimization: comparisons between rationals and constant floats or between
+ integers and constant ratios are now converted to rationals/integers at
+ compile time.
+ * bug fix: problems with NCONC type derivation (reported by Jerry James).
+ * bug fix: EXPT type derivation no longer constructs bogus floating-point
+ types. (reported by Vsevolod Dyomkin)
+ * bug fix: sb-bsd-sockets has correct declaration of the canonname field of
+ addrinfo. (lp#1187041, patch by Jerry James)
+ * bug fix: uninitialized type-error conditions can now be printed.
+ (lp#1184586)
+ * bug fix: tests for sb-bsd-sockets no longer use a predefined port for
+ listening, allowing several tests to run in parallel.
+ * bug fix: during disassembly to *COMPILER-TRACE-OUTPUT* instruction
+ prefixes as used on x86 and x86-64 no longer sometimes print incorrectly.
+ (lp#1085729)
+ * bug fix: Specialised SIMD-PACK types can be negated.
+ * bug fix: Modular arithmetic is more robust. (incidentally fixes another bug
+ reported by Eric Marsden)
+
+changes in sbcl-1.1.8 relative to sbcl-1.1.7:
+ * notice: The implementation of MAP-ALLOCATED-OBJECTS (the heart of
+ ROOM, plus a few SB-INTROSPECT functions) has been completely
+ rewritten; it may now exhibit different bugs than before.
+ * new feature: minimal runtime/compiler intrastructure support for SSE
+ SIMD values, as SB-EXT:SIMD-PACK. Enabled on x86-64 via the build-time
+ feature :sb-simd-pack. This should enable intrinsics extensions as
+ libraries, without patching SBCL itself. Thanks to Alexander Gavrilov
+ for maintaining a branch for so long.
+ * enhancement: RUN-PROGRAM supports a :DIRECTORY argument to set
+ the working directory of the spawned process.
+ (lp#791800) (patch by Matthias Benkard)
+ * enhancement: boxed vectors (vectors of T or of FIXNUM) can now be
+ stack-allocated on PPC.
+ * enhancement: "fixed objects" can now be stack-allocated on PPC.
+ * enhancement: WITH-PINNED-OBJECTS no longer conses on PPC/GENCGC.
+ * enhancement: (sb-introspect:find-definition-sources-by-name x :vop) now
+ also returns VOPs which do not translate any functions, and finds
+ defoptimizer types ir2convert and stack-allocate-result.
+ * enhancement: better type derivation for APPEND, NCONC, LIST.
+ (lp#538957)
+ * enhancement: MAP-ALLOCATED-OBJECTS (the heart of ROOM) now walks the
+ heap in a manner similar to the GC rather than its previous ad-hoc
+ scheme, and is therefore no less and possibly more accurate.
+ * fixes and enhancements related to Unicode:
+ ** the character database information has been updated to Unicode 6.2;
+ ** support for canonical and compatibility decomposition of characters has
+ been added, along with support for primary composition;
+ ** support for Unicode normalization forms of strings (NFC, NFKC, NFD and
+ NFKD) has been included;
+ ** querying the character database for code points not defined by Unicode
+ gives less wrong answers (lp#1178038, reported by Ken Harris)
+ * enhancement: print intermediate evaluation results for some ASSERTed
+ expressions. (lp#789497) (patch by Alexandra Barchunova)
+ * enhancement: x86-64 disassemblies are annotated with unboxed constant
+ values when there are references to (RIP-relative) unboxed constants.
+ * bug fix: type derivation for LOG{AND,IOR,XOR} scales linearly instead
+ of quadratically with the size of the input in the worst case.
+ (lp#1096444)
+ * bug fix: handle errors when initializing *default-pathname-defaults*,
+ sb-ext:*runtime-pathname*, sb-ext:*posix-argv* on startup, like character
+ decoding errors, or directories being deleted.
+ * bug fix: Loading a core with a discontiguous dynamic space now correctly
+ computes the amount of dynamic space used.
+ * bug fix: disassembler missing ",8" on SHLD
+ * bug fix: backtracing through INTERNAL-ERROR signal handlers on systems
+ that provide an "invalid" stack frame link for the signal handler
+ itself now use the saved-fp-and-pc mechanism to pick up from the stack
+ frame in the interrupt (signal) context. This is known to affect
+ threaded FreeBSD/x86-64.
+ * bug fix: some LOOP statements couldn't be compiled.
+ (lp#1178989)
+ * bug fix: sb-sequence:dosequence works on literal vectors.
+ * bug fix: errors in generic arithmetic show the assembly routine's
+ caller on x86 and x86-64. (lp#800343)
+ * bug fix: Compile-time type errors should never result in COMPILE-FILE
+ failure. (lp#943953)
+ * bug fix: Known bad local calls do not cause strange failures when
+ emitting or dumping code. (lp#504121)
+ * bug fix: Multiply-inlined structure constructor don't cause IR2-time
+ codegen errors: type checks are inserted as necessary. (lp#1177703)
+ * bug fix: Unsigned modular arithmetic is correctly converted into signed
+ modular arithemtic (mostly to exploit fixnum-width VOPs). (lp#1026634)
+ * bug fix: a combination of inlined local function with &optional and
+ recursion no longer causes undescriptive compiler errors. (lp#1180992)
+ * bug fix: sub-word BOOLEAN alien types now disregard higher order bits
+ when testing for non-zero-ness.
+ * bug fix: (CONCATENATE 'null ...) no longer fails for generic sequences.
+ (lp#1162301)
+ * bug fix: Type mismatch for the value of conditional expressions are
+ correctly reported when detected at compile-time, instead of complaining
+ about a constant NIL (similar for non-EQ-comparable catch tags).
+ * bug fix: Referring to INLINE global functions as values should not result
+ in a compilation failure. (lp#1035721)
+ * optimization: faster ISQRT on fixnums and small bignums
+ * optimization: faster and smaller INTEGER-LENGTH on fixnums on x86-64.
+ * optimization: On x86-64, the number of multi-byte NOP instructions used
+ for code alignment is now always minimal.
+ * optimization: On 64-bit targets, the compiler generates much faster
+ code for type checks for types known at compile time that are smaller
+ than (SIGNED-BYTE 64) or (UNSIGNED-BYTE 64) and larger than FIXNUM, and
+ their COMPLEX variants.
+ * optimization: On x86 targets, more uses of ALIEN-FUNCALL during cross
+ compilation now inline the INVOKE-WITH-SAVED-FP-AND-PC dance.
+ * optimization: ROOM no longer conses so egregiously on non-x86oid
+ systems.
+ * optimization: associative bitwise operations reassociate patterns like
+ (f (f x k1) k2) to expose (f x (f k1 k2)). Same for + and * of
+ rational values.
+ * optimization: quasiquote expressions now perform more constant folding,
+ instead of consing equal lists at runtime. (lp#1026439)
+ * optimization: local call analysis of inlined higher-order function
+ should converge more quickly, resulting in better code for complex
+ functions.
+ * optimization: On x86-64, medium (word-sized but wider than 32 bits)
+ integer constants are handled more cleverly, especially when they
+ can be represented as sign-extended (signed-byte 32). (Based on a
+ patch by Douglas Katzman)
+ * optimization: IF/IF conversion should reliably result in sane code
+ when (some of) the results of the inner IF are always false or
+ always true.
+ * optimization: On x86oids, variable right shifts of machine words are
+ compiled into straight SAR/SHR, instead of going through the generic
+ VOP. (lp#1066204)
+
+changes in sbcl-1.1.7 relative to sbcl-1.1.6:
+ * enhancement: TRACE :PRINT-ALL handles multiple-valued forms.
+ (lp#457053)
+ * bug fix: :allocation slot option works for condition slots
+ (lp#1049404)
+ * bug fix: redefining conditions does not lead to multiple evaluations of
+ hairy slot initfunctions anymore (lp#1164969)
+ * bug fix: CLASS-DIRECT-DEFAULT-INITARGS now works for condition classes
+ (lp#1164970)
+ * bug fix: function constants now work as initforms and default initarg
+ values of conditions (lp#539517)
+ * bug fix: svref/(setf svref) on symbol macros don't crash the compiler
+ anymore. (Minimal test case provided by James M. Lawrence on sbcl-devel)
+ * bug fix: no more bogus ## references when pretty printing backquoted
+ forms with non-trivial structure sharing. (lp#1161218)
+ * bug fix: Don't signal an error on (setf (documentation nil 'function)
+ "doc"). Signal a style-warning instead when trying to set documentation
+ of NIL for all other documentation types. (Reported by Zach
+ Beane; regression since 2e52fa05.)
+ * bug fix: modular arithmetic involving large constants and conditionals
+ should no longer result in spurious dead code elimination. (Reported by
+ Eric Marsden).
+ * bug fix: our mach exception handler can seemingly called very early in
+ the program execution process on OS X 10.8.0. Try and handle that case
+ robustly, without potentially leaking mach ports too much.
+ * optimization: LOGBITP and LOGTEST optimizations from x86 ported to x86_64.
+
+changes in sbcl-1.1.6 relative to sbcl-1.1.5:
+ * enhancement: the continuable error when defknown-ing over extant
+ fndb entries can be ignored by passing :overwrite-fndb-silently t
+ as a keyword argument to sb-c:defknown (after attributes). Useful
+ to allow defknown to be re-loaded. Use with :allow-other-keys t
+ for backward compatibility.
+ * optimization: compiler is much faster in compiling SVREF and (SETF SVREF)
+ forms.
+ * bug fix: Prevent a make-array transform from modifying source forms
+ causing problems for inlined code. Thanks to Bart Botta.
+ (regression since 1.0.42.11-bis)
+ * bug fix: clear-output calls the correct gray stream routine,
+ sb-gray:stream-clear-output. (lp#1153257)
+ * bug fix: an error is signalled for an invalid format modifier: ~<~@>.
+ (lp#1153148)
+ * bug fix: Better error messages for package operations (lp#1154776)
+ * bug fix: delete-package on a nonexistent package should signal a cerror.
+ (regression since 1.0.37.44).
+ * bug fix: accessing &MORE (stack allocated &REST) arguments checks bounds.
+ (lp#1154946, lp#1072112)
+ * bug fix: compiling make-array no longer signals an error when the
+ element-type is an uknown type, a warning is issued instead.
+ Thanks to James Kalenius (lp#1156095)
+ * bug fix: SEARCH on generic (non-VECTOR non-LIST) sequence types no longer
+ produces wrong results for some inputs. (Thanks to Jan Moringen.)
+ (lp#1153312)
+
+changes in sbcl-1.1.5 relative to sbcl-1.1.4:
+ * minor incompatible change: SB-SPROF:WITH-PROFILING no longer loops
+ by default.
+ * new feature: package local nicknames. See manual for details.
+ * new feature: SB-EXT:MAP-DIRECTORY provides a powerful interface for
+ directory traversal: it is the backend used by SBCL for CL:DIRECTORY.
+ * enhancement: easier to use restarts for resolving name-conflicts
+ resulting from IMPORT, EXPORT, or USE-PACKAGE.
+ * enhancement: variant DEFPACKAGE forms now signal a full error with
+ restarts provided for resolving the situation. (lp#891351)
+ * enhancement: by setting SB-EXT:*ON-PACKAGE-VARIANCE* appropriately variant
+ DEFPACKAGE forms can now signal a full error with restarts provided for
+ resolving the situation. See manual for details. (lp#891351)
+ * enhancement: make-random-state now uses CryptGenRandom as a seed on Windows.
+ (Thanks to Anton Kovalenko.) (lp#1102748)
+ * enhancement: backtrace improvements
+ ** secondary CLOS dispatch functions have better debug names (lp#503081)
+ ** easier to read method names in backtraces. See
+ SB-DEBUG:*METHOD-FRAME-STYLE*.
+ ** SB-DEBUG:PRINT-BACKTRACE and SB-DEBUG:LIST-BACKTRACE are available as
+ forwards-compatible replacements for SB-DEBUG:BACKTRACE and
+ SB-DEBUG:BACKTRACE-AS-LIST.
+ ** SB-DEBUG:*SHOW-ENTRY-POINT-DETAILS** has been deprecated, as the same
+ information is available in less intrusive form as frame annotations.
+ * enhancement: SB-POSIX now provides MAP-ANON.
+ * enhancement: test-suite results are colorized, failures in red, unexpected
+ success in green. Works on Windows and on terminals with ANSI escape code
+ support. Can be disabled with --no-color.
+ * optimization: SB-CONCURRENCY:QUEUE implementation is more efficient.
+ (thanks to James M. Lawrence)
+ * bug fix: no more unused variable style warnings from RESTART-CASE
+ macroexpansion (lp#1113859)
+ * bug fix: no more unused variable style warnings when loading
+ sb-bsd-sockets.asd (lp#1114681)
+ * bug fix: deleting a package removes it from implementation-package
+ lists of other packages.
+ * bug fix: SB-SPROF:WITH-PROFILING is now usable in the Slime REPL on Darwin.
+ This does not fix the occasional "interrupt already pending" issue, though.
+ * bug fix: (setf (documentation 'x 'function)) and
+ (setf (documentation #'x t)) set documentation in different places.
+ (regression since 1.0.43.63)
+ * bug fix: build on newer glibc. (lp#1095036)
+
+changes in sbcl-1.1.4 relative to sbcl-1.1.3:
+ * optimization: LOOP expressions using "of-type character" have slightly
+ more efficient expansions.
+ * bug fix: very long (or infinite) constant lists in DOLIST do not result
+ in very long compile times or heap exhaustion anymore. (lp#1095488)
+ * bug fix: `#3(1) is read as #(1 1 1), not as #(1). (lp#1095918)
+ * bug fix: adjust-array ignored :initial-element for simple-vectors.
+ (lp#1096359)
+ * bug fix: optimizations to MAKE-INSTANCE with literal list initargs no
+ longer cause infinite loops (on circular data) or violate eqlity
+ constraints. (lp#1099708)
+ * bug fix: FIRST was not being open coded. (regression from 1.1.0)
+
+changes in sbcl-1.1.3 relative to sbcl-1.1.2:
+ * enhancement: warnings about bad locale settings, LANG, LC_CTYPE, etc.
+ (lp#727625)
+ * enhancement: support for C-c to interrupt the foreground thread
+ on Windows. (Thanks to Anton Kovalenko.)
+ * enhancement: STDCALL alien callbacks. (Thanks to Anton Kovalenko.)
+ * enhancement: Safepoint builds on POSIX platforms can now optionally be
+ built without pseudo-atomic sequences and their run-time overhead.
+ * enhancement: Threads created outside of Lisp can enter Lisp through
+ alien callbacks and appear as Lisp threads for the duration of that
+ function invocation. On safepoint builds only.
+ * enhancement: Miscellaneous improvements to namestrings and underlying
+ calls to OS functions for file system access on Windows. (Thanks to
+ Anton Kovalenko.)
+ * enhancement: The MSI installer support for Windows now uses Windows
+ Installer XML at least version 3.5 and includes various usability
+ improvements. (Thanks to Dmitry Kalyanov and Anton Kovalenko.)
+ * enhancement: The sb-bsd-sockets contrib now supports non-blocking-mode
+ on Windows. (Thanks to Anton Kovalenko.)
+ * enhancement: The Windows backend now supports the x86-64 platform.
+ (Thanks to Anton Kovalenko.)
+ * bug fix: fasls are now once again directly executable (on platforms
+ supporting shebang lines, with a suitably-installed sbcl).
+ * bug fix: make.sh --help no longer runs clean.sh. (lp#937001)
+ * bug fix: Fix CAS access to slots of direct instances of structure classes
+ in the presence of subclasses sharing the same conc-name.
+ * bug fix: Logical pathname namestrings on Windows have been changed to
+ be lower-case, to minimize differences between POSIX and Windows.
+
+changes in sbcl-1.1.2 relative to sbcl-1.1.1:
+ * notice: System requirements for SBCL on Microsoft Windows: Windows NT 5.1
+ or newer (Windows XP, Server 2003) is required. Support for Windows 2000
+ (NT 5.0) is no longer being maintained.
+ * notice: Starting with this version, SBCL on Windows no longer supports
+ building with disabled thread support.
+ * enhancement: frlocks have been added to SB-CONCURRENCY contrib module.
+ * enhancement: New feature sb-dynamic-core allows the runtime to be
+ rebuilt or relocated without requiring changes to the core file on
+ all linkage table platforms. Required on Windows. (Thanks to Anton
+ Kovalenko.)
+ * enhancement: Console I/O streams use UCS-2. (Thanks to Anton Kovalenko.)
+ * enhancement: I/O on Windows unnamed pipes is interruptible
+ asynchronously using interrupt-thread, timers when running on Windows NT
+ version 6.1 or newer (Windows 7, Server 2008 R2). (Thanks to Anton
+ Kovalenko.)
+ * enhancement: Support for the experimental safepoint-based stop-the-world
+ protocol on the PowerPC platform.
+ * bug fix: Non-blocking reads from the Windows console were not necessarily
+ non-blocking. Thanks to Anton Kovalenko. (lp#308923)
+ * bug fix: stability of threads on Windows has been improved upon through
+ an updated stop-the-world protocol (thanks to Anton Kovalenko).
+
+changes in sbcl-1.1.1 relative to sbcl-1.1.0:
+ * enhancement: WITH-COMPILATION-UNIT no longer grabs the world-lock.
+ (COMPILE and COMPILE-FILE still do.)
+ * optimization: the SPARC backend now supports the precise generational
+ (GENCGC) garbage collection. Enabled by default on Solaris/SPARC and
+ Linux/SPARC. Thanks to Raymond Toy (via CMUCL).
+ * enhancement: add experimental support for the SB-THREAD feature and the
+ timer facility on Windows. Thanks to Dmitry Kalyanov and Anton Kovalenko.
+ Threads are enabled by default, and this version of SBCL is considered
+ to be the last and final release to officially support building with
+ threads disabled.
+ * optimization: The compiler no longer rotates loops in some cases where
+ this transformation actually lead to worse code being generated.
+ * bug fix: SB-CLTL2:MACROEXPAND-ALL correctly handles shadowing of
+ symbol-macros by lexical bindings.
+ * bug fix: stack allocation was prevented by high DEBUG declaration in
+ several cases.
+ * bug fix: SB-EXT:GC-LOGFILE signaled an error when no logfile was set.
+ (thanks to SANO Masatoshi)
+ * bug fix: PARSE-NATIVE-NAMESTRING performed non-native parsing when
+ :JUNK-ALLOWED was true.
+ * bug fix: type derivation inferred overly conservative types for
+ unions of array types. (lp#1050768)
+
+changes in sbcl-1.1.0 relative to sbcl-1.0.58:
+ * enhancement: New variable, sb-ext:*disassemble-annotate* for controlling
+ source annotation of DISASSEMBLE output. Defaults to T.
+ * enhancement: TIMEOUT arguments added to WITH-MUTEX and WITH-RECURSIVE-LOCK, and
+ WAIT-P argument added to WITH-RECURSIVE-LOCK.
+ * enhancement: SB-EXT:ATOMIC-PUSH and SB-EXT:ATOMIC-POP allow atomic operations
+ on list heads.
+ * enhancement: Optional features (not enabled by default) allow the
+ use of signals for inter-thread synchronization to be reduced on certain
+ supported platforms (currently Linux, Solaris, and FreeBSD on x86 and
+ x86-64). Set (and :sb-thread :sb-safepoint :sb-thruption :sb-wtimer)
+ to test these experimental features. Known remaining bugs include minor
+ thread safety issues, less timely freeing of memory by GC, and certain
+ (not yet optimally low) runtime overhead. Thanks to Anton Kovalenko.
+ * optimization: CL:SORT and CL:STABLE-SORT of lists are faster and use fewer
+ comparisons, particularly on almost-sorted inputs.
+ * bug fix: Reading floats with large exponents no longer takes too much time
+ before reporting that the exponent is too large.
+ * bug fix: SB-BSD-SOCKETS:SOCKET-RECEIVE with a UDP socket now works
+ correctly when the received datagram is larger than the provided buffer.
+ (lp#1023438, thanks to Robert Uhl)
+ * bug fix: SB-EXT:GET-CAS-EXPANSION returned a bogus read-form when given
+ a SYMBOL-VALUE form with a constant symbol argument.
+ * bug fix: SB-EXT:GET-CAS-EXPANSION signaled an error when a macro expanding
+ into a DEFCAS defined place was used as the place.
+ * bug fix: FIND and POSITION signaled a type-error when non-bits where looked
+ for from bit-vectors.
+ * bug fix: a race condition around thread creation could (in SBCL 1.0.57)
+ lead to internal errors or crashes (lp#1058799).
+ * documentation: a section on random number generation has been added to the
+ manual. (lp#656839)
+
+changes in sbcl-1.0.58 relative to sbcl-1.0.57:
+ * enhancement: implicit generic function warnings now specify the package
+ in which the new generic function is being created.
+ * enhancement: SB-EXT:ATOMIC-UPDATE makes it easy to perform non-destructive
+ updates of CAS-able places (similar to Clojure's swap!).
+ * enhancement: run-program no longer decodes and re-encodes environment when
+ :environment argument is not provided. (lp#985904)
+ * enhancement: errors during compiler-macro expansion no longer cause
+ runtime errors, only a compile-time warning, otherwise behaving as if
+ the compiler macro had declined to expand.
+ * optimization: On x86-64, code alignment of block headers is done with
+ multi-byte NOPs now instead of repetitions of the single-byte NOP.
+ * optimization: MAP-INTO is substantially faster when the target sequence is
+ of unknown type; mapping into lists is no longer O(N^2). (lp#1001043,
+ thanks to James M. Lawrence)
+ * optimization: the compiler no longer heap-conses to check exits in cases
+ where the exit function is dynamic extent, or when it can prove the exit
+ function cannot escape. (lp#1002534)
+ * optimization: SB-SEQUENCE:DOSEQUENCE is faster on vectors of unknown
+ element type, and vectors that aren't SIMPLE-ARRAYs.
+ * optimization: CL:SORT and CL:STABLE-SORT are more efficient in execution
+ speed (around 1/3 the time in some cases), and a little better in terms of
+ comparison calls. (Thanks to Takeru Ohta)
+ * bug fix: On SPARC, a limitation on the number of code constants emittable
+ by the compiler has been lifted, allowing certain long functions to
+ compiled and assembled which had previously been unsupported; fixes
+ cl-bench on this ISA (lp#1008996).
+ * bug fix: potential for infinite recursion during compilation of CLOS slot
+ typechecks when dependency graph had loops. (lp#1001799)
+ * bug fix: error forms reported with some program-errors were not escaped
+ properly.
+ * bug fix: functions from EVAL are now on more equal footing with functions
+ from COMPILE. (lp#1000783, lp#851170, lp#922408)
+ * bug fix: ENSURE-GENERIC-METHOD-COMBINATION accepts method combination
+ objects as its :METHOD-COMBINATION argument, not just lists designating
+ method combinations. (lp#936513)
+ * bug fix: run-program no longer unconditionally uses /tmp/ for temporary
+ files. (lp#968837).
+ * bug fix: restore build on solaris/sparc. (lp#1008506)
+ * bug fix: an issue with LDB in the PowerPC backend has been resolved;
+ this fixes an issue found with cl-postgres (thanks to Tomas Hlavaty).
+ * bug fix: compiler-macro lambda-lists specifying non-keyword symbols
+ as keyword arguments no longer accidentally match unevaluated symbols
+ against them.
+ * bug fix: FORMAT used to loop infinitely in some cases when a COLINC
+ parameter was zero, now it signals an error. (lp#905817, fixed since
+ 1.0.56.19)
+ * bug fix: run-program with :pty t no longer makes the pty as the process's
+ controling terminal.
+ * bug fix: spawning threads on OS X 10.8 no longer crashes the process
+ (lp#1012811).
+
+changes in sbcl-1.0.57 relative to sbcl-1.0.56:
* RANDOM enhancements and bug fixes:
** bug fix: the range and distribution of random integers could be
catastrophically wrong when the compiler derived the type of its
** bug fix: the distribution of random integers is now completely
uniform even when the specified limit is not a power of two.
(Previously some values could be about 0.1 % more probable than
- others in the worst case.)
+ others in the worst case.) (lp#309467)
** RANDOM on large integer arguments is generally faster and conses
less than before; this is visible for fixnums above a length of
about 24 bits, but extremely so for bignums: the old implementation