Release notes for CMU Common Lisp 17f, 4 November 94 17f contains bug fixes, PCL enhancements, and support for three new platforms: Sun SPARC/Solaris SGI MIPS/Iris (no load-foreign) DEC Alpha/OSF1 (no Motif or load-foreign) The HPPA/HPux and Sun/Sunos 4.x platforms are still available. Distributions are now gzip'd, not compressed. Other highlights: Motif, load-foreign on Hp/Ux, image size improvements, MIT loop, some documentation of some previously undocumented features. [Also see the README file for platform-specific notes.] Bug fixes to basic CL support: -- Fixed ADJUST-ARRAY to not flame out on arrays containing a zero-length dimension (hence having no elements whatsoever.) -- In SIGNAL, bind *break-on-signals* to NIL even before doing the type test so that we don't wedge the error system if the type is malformed or undefined. -- Fixed bugs with EOF handling in READ-LINE. -- In DEFINE-CONDITION, don't warn about probable error unless both initarg and initform are missing. -- In OPEN, fixed :direction :io :if-does-not-exist :create to actually create. -- Fix problem in LOAD-FOREIGN where (especially on SunOS), the failure to page-align loaded code caused errors when load-foreign was done more than once. -- In OUTPUT-INSTANCE, check for the layout being invalid before calling the print function, since the print function might error. -- Closing composite streams (broadcast, etc.) no longer closes the component streams. -- Fixed pprint goof that didn't actually break anything, but wasted effort. -- (COERCE x 'FLOAT) now convert to a single-float (instead of being an error.) Also, we now check that numeric coercions actually return a value of the specified type (which they might not if the type is not a symbol, e.g. (coerce 0 '(complex float)). Possibly these should "do the right thing", but it seems better to error than quietly do the wrong thing. -- Fixed a bug in FLOAT-BIGNUM-RATIO that could cause an assertion failure when floating particular bignum ratios (or possibly reading particular float values.) Miscellaneous enhancements: -- LOOP is now the MIT/Symblics loop, so whatever it does is by definition correct and The Right Thing. -- PURIFY is now exported as EXT:PURIFY. This function can greatly improve the GC performance of many large programs by moving code and static data into non-collected storage. This is a "poor man's generational GC". Environment compaction now done by purify. -- Some reduction in the size of the image (and of GC scanned memory) from tweaks to build process. -- Binary input can now be done from string streams (from David Axmark.) -- Debugger no longer aborts printing of a frame when printing one arg gets an error (from Harris.) -- LOAD-FOREIGN support for HP/Ux (from TSM.) -- Add sap-ref-64 (only on Alpha). -- Changes to EVAL, ROOM and site-init to work better in a runtime core image (without the compiler loaded.) -- *BEFORE-SAVE-INITIALIZATIONS* is now actually done before saving. Compiler: -- Fixed some problems with multiple values and cleanup code in byte compilation. Also added EXT:*COMPILE-PROGRESS* printout. -- Fixed some problems with internal errors when a function was just declared to be FUNCTION. -- Now allows stream args to compile-file and doesn't attempt to constant-fold pathname functions (which depend on *default-pathname-defaults*.) -- Fixed a case where dead local function in top-level code could cause an internal error. -- Fix compiler-macro expansion to correctly handle macros that pass (by returning the unmodified form.) -- Fix spelling of :COMPILE-TOPLEVEL and :LOAD-TOPLEVEL in EVAL-WHEN. -- If compile-file is :block-compile T, the entire file is block-compiled as a unit (even if it contains embedded START-BLOCK/END-BLOCK declarations.) -- Virtually all of the compiler is now compiled without type checking, giving some space and speed benefit. CLX: -- Merged with CLX R5.02 (no substantive changes). -- In read-resources, trim off spaces, tabs, and "'s in #include file name -- If CLX is compiled when PCL is loaded (as is now done in the binary distribution), DRAWABLE, WINDOW and PIXMAP will be defined as PCL classes (which can be subclasses.) This is compatible with various CLX-based toolkits. -- Fix some CONS declarations to be LIST because they aren't conses on the last iteration (when the body isn't executed.) -- Fix incorrect slot type declaration for DISPLAY-AUTHORIZATION-DATA. -- Changed holding-lock not to turn off GC, which could cause event handlers and other code to run for ling periods of time without garbage collecting. Instead we bind all the GC hooks to (), since it was their invocation that could cause bad recursive entry of CLX. Hemlock: -- Fixed problem in Hemlock recursive error handler (hadn't been properly updated for ANSI conditions.) -- Add ignore handler for client-message events. -- Deleted some setting of hi::*hack-hunk-replace-line* to T, since we may want it explicitly disabled in the init file. -- Dylan mode now infix-oriented. Motif interface: -- Fixed a bug in the generation of Interface.h which was preventing motifd from being successfully compiled on HP/Ux, Solaris, ... -- use pcl::precompile-random-code-segments to minimize run-time compilations. -- Add INVOKE-TTY-DEBUGGER. Don't try to invoke motif debugger if we didn't succeed in opening a connection. -- Print warning in status hook when server dies. -- Made server to fflush after all output so that we see motifd output when it is run on a pipe. -- Fixed severely broken INSPECT-CLOS-PANE according to patch from Marco Antoniotti. -- Fix from Marco Antoniotti to actually remove handlers from the table in remove-event-handler. -- Fix to TOOLKIT-WRITE-VALUE to allow it to write either signed or unsigned word integers. -- Improved error recovery and internal error reporting. PCL: -- Structure-object is now no longer shadowed in PCL. Code that was using PCL::STRUCTURE-OBJECT will now work better. -- BUILT-IN-CLASS, CLASS-NAME, CLASS-OF and FIND-CLASS are once again exported from PCL. This will cause a name conflict if anyone use-package's PCL, but this will at least warn about the distinction. Probably you shouldn't USE-PACKAGE PCL for this reason, but you can now say PCL:FIND-CLASS instead of PCL::FIND-CLASS. It is also possible to use SHADOW or SHADOWING-IMPORT to resolve the conflict. -- Fix to update-instance-for-different-class. -- When updating arg-info slots, check to see if the value is already there. This can reduce non-shared pages. -- Improved handling of invalid structure instances. -- Fix a problem with PCL clobbering byte functions when setting their names. -- New parameterized version of use-dispatch-dfun-p which avoids pessimizing GFs with many methods. -- Fix to :after methods on accessor functions. Also, fixed some problems with the result of get-secondary-dispatch-function1 when there are no methods. -- Add compiler-macro for pcl:find-class which does lookup at load-time when the name is a constant. -- Definitive tweak for handling function-p in compute-secondary-dispatch-function1 which avoids an infinite recursion. -- When signalling an incompatible superclass error, added a hint to the message to check out VALIDATE-SUPERCLASSES. Lisp code: -- Fixed Sparc GC bug fix (L2 never scavenged.) -- On all non-Mach platforms, changed the default for CMUCLLIB to be /usr/local/lib/cmucl/lib. -- On SunOS, added "dynamic segments" patch which prevents the "out of segments" errors that could happen when a Lisp memory management table overflowed. Build tools: -- Fix compilation of motif interface to actually generate the C header files Interface.h, etc. -- Some changes to reduce compiler warnings -- In compile-all, -clean and -noupdate have been flushed. -clean is now done by the clean-build script. -- Add some scripts from David Axmark that automate the entire build process: tools/variant-lisp tools/build-and-install Release notes for CMU Common Lisp 17e, 11 March 94 17e is primarily a bug-fix release. This release is also available on Hewlett-Packard 700-series workstations running HP/UX version 9.x. The main missing feature on HP/UX is LOAD-FOREIGN. BASIC COMMON LISP SUPPORT CODE: Enhancements: -- The function EXT:PURIFY is now exported. This was formerly only documented as usable via the :PURIFY argument to SAVE-LISP. Great improvements in GC runtime and memory use can be obtained when PURIFY is called after loading in large amounts of code or creating large datastructures that are unlikely to become garbage. -- EXT:PURIFY (and thus (EXT:SAVE-LISP ... :PURIFY T) now compact the hashtables representing the global namespace. Formerly this feature was undocumented and only used during the initial system build. -- There is now a "runtime" version of the Lisp image which omits the compiler and loads some other code byte compiled, giving a 60% size reduction. See README file, "Runtime distribution" section. -- Changed the byte-function and byte-closure funcallable-instances to use &more args instead of a &rest arg (eliminating the main source of grautitous consing in byte code.) -- Allow pretty-printer to work without compiler loaded. -- Improved (and thus changed) the SXHASH of lists and other composite objects. -- Added to *FEATURES* all of the ANSI CL features that seemed plausible. Any non-ANSI compliance is now a bug, not a feature... -- Picked up Miles' changes to apropos and apropos-list that allows one to supply NIL for the package, meaning all packages. Bug fixes: -- In SIGNAL, bind *break-on-signals* to NIL even before doing the type test so that we don't wedge the error system if the type is malformed or undefined. -- Fix arg ordering for GETF in new condition support. Default the name slot in DEFMACRO-LAMBDA-LIST-BIND-ERRORS to NIL (instead of leaving it unbound.) -- In READ-N-BYTES, only signal EOF when we read and got 0 bytes, not whenever read returns less than we wanted. Also, fix the case of small reads with an empty buffer not to leave garbled buffer pointers if we have to do more than one read. These bugs mostly affect the X interface and its users (Garnet was having problems.) -- Changed YES-OR-NO-P and Y-OR-N-P to force output. -- Fixed COUNT :TEST-NOT to actually negate the test. -- Fixed COERCE to call type-expand so that people can coerce to deftypes. -- Rename STEP to DEBUG::SET-STEP-BREAKPOINT because step is supposed to be the exported step macro, not some internal routine. -- Fixed DEFPACKAGE to combine multiple use, export, and import options correctly. Fixed the warning about no longer using some packages to actually work. -- Fixed GCD to correctly handle MOST-NEGATIVE-FIXNUM. -- Flush Hemlock font change hack in default site-init.lisp, since it was causing problems. CLOS (PCL): -- Add some hacks for coercing lisp:class objects to pcl:class. -- Export MAKE-LOAD-FORM-SAVING-SLOTS, not MAKE-LOAD-FORM-DUMPING-SLOTS. -- Fix some problems with the LISP:CLASS class-precedence-list suddenly becoming empty when a class is redefined. -- Fixed SIMPLE-STRING to list STRING in the inherits list. -- Fix to update-instance-for-different-class. COMPILER: Enhancements: -- Added a type inference method for LOG's result type. -- Added support for "more args" which are like rest args, but don't cons. This is used for calling into annd out of byte-compiled core and in the PROFILE package. -- Increase *inline-expansion-limit* from 50 to 200. Change "inline expansion count exceeded" warning to be a note. Don't inline expand in interpreted code. Bug fixes: -- Allow stream arg to compile-file. -- Changed assert-definition-type to quietly ignore non function-types. -- Allow stream arg to compile-file. -- Deleted incorrect type declaration in process-1-ftype-proclamation. Deleted test for function-type-p around call to assert-definition-type, since this is now folded into the function. Previously several calls weren't doing this necessary check. -- Fix a problem where spurious argument types could be inferred for optional arguments in byte-compiled code (caused yes-or-no-p to fail.) -- Fixed an internal error related to tail local calls. -- Fixed assertion failure about COMPONENT-NEW-FUNCTIONS with some inline expansions. -- Fixed (TYPEP X '(REAL ...)) to actually test against REAL, not NUMBER. -- Fixed a problem where top-level MULTIPLE-VALUE-CALL forms could cause an assertion falure. -- Don't try to compile flow-graph components that are unreachable. This avoids some internal errors when unreachable local functions are deleted. -- Fix problem with byte-compiling PROGV getting an internal compiler error. EXTENSIONS: Misc extensions: -- Added an executable-only optional to EXT:UNIX-NAMESTRING so that we can expand path:foo into the first executable in the path named foo, not just the first file -- Changed RUN-PROGRAM to only try to run executables, and to flame out if it can't be found in path:. -- Profile: use &more instead of &rest to avoid consing the &rest arg list. Fixed PROFILE:REPORT-TIME to not choke if any of the function names are not symbols (e.g. setf functions). [NOTE: Motif server doesn't work on HP yet.] Motif interface: -- Added an optional STYLE arg to the graphical inspector to make the function consistent with its doc string (as well as the old inspector). -- Tried to make starting the Motif server more robust, or at least give some intelligible error message. -- Added some extra protection against recursive invocation of the windowing debugger. -- Don't try to invoke motif debugger if we didn't succeed in opening a connection. -- Print warning in status hook when server dies. -- Made server fflush after all output so that we see output when it is run on a pipe. Hemlock: -- Add window-buffer-hook for echo area buffer to make absolutely sure we can't change the eacho area window to point to another buffer. -- Give a sensible error message if we can't find the slave utility (e.g. lisp) to run. -- Restored the behavior of scribe-file to cd to the directory of the .mss file before running scribe. -- Change default fonts to be courier, flushing all mention of 8x13. Remove font-setting in default site-init file, since it was causing problems. -- Add FILE-POSITION method for region streams so that the compiler's read error recovery works for "Editor Compile Defun", etc. -- When reading a file, set BUFFER-WRITABLE to T before doing anything so that we don't get spurious read-only errors. Release notes for CMU Common Lisp 17c, 19 September 93 17c is a new major release of CMU Common Lisp. An overview of changes: -- New structure object representation and class support in the type system. -- PCL better integrated with CMU type system. -- New CLX, PCL. -- Numerous ANSI changes. -- Byte-code compilation option offers more compact code. -- Improvements in compiler source-level optimization, inline expansion and instruction scheduling. -- New fasl file format (you must recompile.) -- Calling of SETF functions is now efficient. -- Speed and space tuning in the compiler and runtime system. -- New TTY debugger commands support stepping compiled code. -- A graphical debugger and inspector based on a Motif interface. -- Changes in the startup code and SAVE-LISP increase portability. -- Preliminary support for the HP Precision Architecture under HP-UX 9.0 (i.e. for HP 700 series.) -- New User's manual. Note that the online Info document has not been updated yet. And of course, bug fixes... Basic runtime code changes: ANSI cleanups: -- These functions are now defined: ARRAY-DISPLACEMENT CELL-ERROR-NAME COMPILE-FILE-PATHNAME COPY-STRUCTURE DELETE-PACKAGE INTERACTIVE-STREAM-P LOAD-LOGICAL-PATHNAME-TRANSLATIONS LOGICAL-PATHNAME LOGICAL-PATHNAME-P LOGICAL-PATHNAME-TRANSLATIONS MAKE-LOAD-FORM MAKE-LOAD-FORM-SAVING-SLOTS MAP-INTO OPEN-STREAM-P TRANSLATE-LOGICAL-PATHNAME -- The conditions FLOATING-POINT-INEXACT, FLOATING-POINT-INVALID-OPERATION, and the type FILE-STREAM are now defined. -- Logical pathnames are now implemented. (See "pathnames" below and the user's manual.) -- Allow :UNSPECIFIC and :WILD in all pathname slots where they should be legal; don't always use a :MULTI-CHAR-WILD pattern. -- Support for :COMPILE-TOP-LEVEL, :LOAD-TOP-LEVEL and :EXECUTE keywords to EVAL-WHEN. -- #S readed no longer forces keywords into the keyword package. -- Various changes to DEFSTRUCT described below. -- IN-PACKAGE now prints a warning if any arguments other than the package name are supplied and signals a correctable error if the package doesn't exist yet. -- DEFPACKAGE now tells you about inconsistencies between any existing package and the DEFPACKAGE form. -- Packages: - DELETE-PACKAGE function added according to X3J13/92 specification. Most operations on deleted packages signal an error. - Changed IN-PACKAGE to conform to the new definition. But if you use an old-style IN-PACKAGE, it will use the old behavior. - Rewrote DEFPACKAGE to tell you about inconsistencies between the package and the DEFPACKAGE form. -- Conditions: - Changed conditions to be non-structure (but also non-PCL) instances that support multiple inheritance and the other required CLOS operations. Removed SIMPLE-CONDITION hacks for simulating multiple inheritance. - DEFINE-CONDITION now corresponds to proposed ANSI CL, which is a subset of DEFINE-CLASS syntax and is incompatible with the earlier KMP syntax. Note that in particular, :INITARG and :READER options are now effectively required on every slot. - SIMPLE-CONDITION-FORMAT-STRING renamed to SIMPLE-CONDITION-FORMAT-CONTROL. - Implemented CONDITION-RESTARTS ANSI cleanup & WITH-CONDITION-RESTARTS macro. This provides a way to say that restarts are relevant only to a certain condition. - Added STYLE-WARNING and PARSE-ERROR conditions. - Added report method for END-OF-FILE and changed system code to signal it. - Added PRINT-NOT-READABLE condition and made people use it. -- Remove some spurious LISP package exports, and add missing ones. Delete package setup for optional subsystems (hemlock, etc.) -- Renamed SPECIAL-FORM-P to SPECIAL-OPERATOR-P. -- Renamed GET-SETF-METHOD-MULTIPLE-VALUE to GET-SETF-EXPANSION and DEFINE-SETF-METHOD to DEFINE-SETF-EXPANDER. The old names are still defined for CLtL1 compatability. -- Added degenerate versions of STREAM-EXTERNAL-FORMAT, FILE-STRING-LENGTH and the :EXTERNAL-FORMAT argument to OPEN. -- Hash-table code largely rewritten. MAKE-HASH-TABLE now conforms to the X3J13 spec. Hash-tables can now be dumped as constants in fasl files. Bug fixes: -- Made the "modules:" search-list (used by REQUIRE) default to the current directory. -- Do a BOUNDP check so that references to undefined types inside of a WITH-COMPILATION-UNIT but outside of the compiler won't cause undefined-variable errors. -- Set up a default for modules: search-list. -- Changed BACKQ-UNPARSE to check for improper lists instead of getting an internal error. Some meaningless backq forms will now pprint as: "### illegal dotted backquote form ###". -- Added SIMPLE-STYLE-WARNING, and spiffed up the simple-condition hacks so that (typep x 'simple-condition) works. -- In LOAD-FOREIGN, use unix-namestring on each file before passing it to the linker to get rid of search lists. -- Fixed the printer to stop at the fill pointer for strings with fill pointers. -- Fixed PPRINT-DO to not flame out if one of the binding lists is NIL. -- Fixed load to not always consider files with NIL type to be source files. If the file exists as specified, then look at the header instead of trying to default the type. If :CONTENTS is specified, then don't try defaulting types. -- Fixed FORMAT-EXP-AUX to correctly handle variable width fields when the argument is negative. -- Use ~C instead of ~A when printing float exponent marker so that *PRINT-READABLY* doesn't mess things up. -- Fixed CLEAR-INPUT on file descriptor streams to flush any unread chars. -- Now that +0.0 and -0.0 are no longer EQL, fixed ATAN to deal with them correctly. -- Changed SAVE-LISP to pad the core file out to CORE_PAGESIZE bytes, so that when we mmap it back it, we won't get bus errors if the real page size is less then the CORE_PAGESIZE. -- Really really fixed GET-SETF-METHOD-MULTIPLE-VALUE for local macros. Also, in the recursive calls, people were not propagating the environment through, and in some places were not recursing with the multiple-value version. -- Fixed FLOAT to float ratios precisely by using integer division instead of float division. This fixes a problem where a bit or two was lost on READing floats. -- In the Alien type= method for FUNCTION, call ALIEN-TYPE-= instead of calling ALIEN-TYPE-P with two args. [Fix from Mike Clarkson.] Compiler changes: Enhancements: -- If the argument to compile-file is already absolute, then don't bother expanding it into the TRUENAME. This allows search-lists or logical pathnames to be preserved in the defined-from info (for DESCRIBE, "Edit Definition", etc.) -- Signal a compile-time error for division by constant 0. -- Replaced the FORMAT transform with one that uses FORMATTER for more complete handling of format directives. This is only enabled when speed > space. -- Compilation to a dense byte-code is now supported, see below. -- Semantic analysis/optimization of function calls has been revamped so that optimizations are done more consistently, especially when the call is a funcall. -- A new approach is now taken to inline expansion, allowing inlining to be done in more cases. In particular: - local functions from LABELS or block compilation can now be inlined, - global function definitions made inside of a local macro or special declaration can now be inlined. Inline expansion is now divided into two separate parts: - Semi-inline expansion introduces a local definition of a global function that has an expansion available. This is now exactly equivalent to block-compiling together with that DEFUN. - Local call analysis introduces new copies of locally defined INLINE functions. This duplication is limited by EXT:*INLINE-EXPANSION-LIMIT* (default 50) to prevent indefinite expansion of recursive functions. This limit may need to be increased for compilations containing many legit inline expansions in order for all calls to be inlined. -- SETF functions are now better supported. Calling a compile-time constant SETF function is now just as efficient as calling a function named by a symbol. This is done by resolving function names to "fdefinition objects" at load time. SYMBOL-FUNCTION of a non-constant symbol is now somewhat slower, since the fdefinition must be located by a table lookup. -- Assembler and disassembler have been reimplemented yet again, giving improved portability and scheduling. -- Assembly optimization is now enabled, giving large speed/space improvements on MIPS and some on SPARC. This optimization is done when speed > compilation-speed (i.e. not by default) since it significantly slows compilation. -- [mips] Lots of tweeks in order to use NIL and 0 directly from the registers holding them instead of copying them into a new register and then using it. -- New funcallable-instance support (for PCL, etc.) Now funcallable instance functions must be specially compiled, which is indicated by creating them with KERNEL:INSTANCE-LAMBDA. -- Moved assumed-type recording of unknown function calls from the beginning of IR1 to the end so that we have the best type information about the arguments. -- Generalized the static-mumble-offset routines to also consider nil a static symbol at offset 0. ANSI changes: -- Make compiler error functions use the condition system. This ANSI cleanup has two advantages: 1] compiler-warning and warn are now equivalent, so uses of WARN will be counted in the warning total and given source context. 2] user handlers can be defined to notice or suppress output. -- Made DYNAMIC-EXTENT declaration recognized & ignored. -- Allow non-keyword keyword names when the &key keyword is specified separately. In FUNCTION and VALUES types, allow non-keyword symbols; you must now explicitly the ":" in order to get the usual keyword naming. -- Compiler-macros are now supported. See DEFINE-COMPILER-MACRO. -- Minor tweeks to conform to X3J13 cleanup MACRO-DECLARATIONS:MAKE-EXPLICIT. -- Fixed SYMBOL-MACROLET to allow declarations as per X3J13 cleanup SYMBOL- MACROLET-DECLARE:ALLOW. When declaring things about symbol macros, type declarations just wrap (the type ...) around the expansion, special declarations signal an error, and ignore/ignorable declarations are ignored. Tuning: -- Tuning based on instruction-counting profiling of the compiler --- added missing declarations, etc. This also motivated some general array/sequence optimizations (see below.) -- Added block compilation declarations. Moved some stuff around to get better locality. -- Changed IR1-ERROR-BAILOUT to do fewer special bindings. -- Inline expand some simple utility functions. -- Some changes to increase the sharing among some of the automatically generated functions in the compiler backend. Bug fixes: -- [sparc,mips] Fixed a bug in default-unknown-values where it wouldn't default the first unsupplied value to nil if more then 6 values where supplied. -- Structure slot accessors are no longer constant-folded, because that was causing problems with some MAKE-LOAD-FORM hacks. -- Fixed numeric type inference to realize that contagion causes the result of SINGLE-FLOAT & REAL to be FLOAT, not SINGLE-FLOAT. -- In FINALIZE-XEP-DEFINITION, if not the current global definition, just leave the defined type alone, instead of clobbering it with FUNCTION. A benefit of this is that COMPILE doesn't trash the function type. -- Don't run the back-end(s) on components with no code. -- Don't compile load-time-value lambdas if they've already been compiled because they ended up in a non-top-level component. Also, the function holding a load-time-value form also has a more sensible debug name. -- Fixed a problem with ASSERT-DEFINITION-TYPE when we have a keyword arg with a non-constant default. -- Fixed several uses of FIND to use EQUAL instead of EQL to compare function names, because (SETF mumble) is now a valid function name, and isn't necessarily EQL. -- Bind *GENSYM-COUNTER* around compilation that it doesn't get side effected. -- Fixed a bug in type inference which seems to have generally prevented anything from being inferred about function result types. -- Fixed several bugs related to the handling of "assignment" local functions (that correspond to a tail-recursive loop.) -- Added a hack to IF-IF optimization to hopefully prevent some spurious unreachable code notes. -- Fixed call to CONTINUATION-DERIVED-TYPE to be CONTINUATION-TYPE so that we don't choke on values types in the functional position. -- Fixed the handling of +/- 0.0: - = is no longer converted to EQL, but is directly handled by the backend. - EQL is converted into a raw comparison of the bits. -- Fixed several bugs which caused fatal compile-time errors. -- Fixed LET* to correctly use the internal policy (not the interface policy) for all bindings, not just the first. -- In local call VOPs, must load CALLEE-NFP with MAYBE-LOAD-STACK-TN, since it might not be in a register. -- When iterating over the lamdba-calls in unconverting tail calls, have to ignore any deleted lambda. -- Fixed listify-rest-arg. It was leaving a tagged pointer to unallocated memory in a descriptor register, which would confuse the garbage collector if this value was still around. -- Compile and dump package manipulation forms before evaluating them, so they are dumped with respect to the state of the package system before the form was read, not after. Sparc: -- New pseudo-atomic speeds allocation. -- Added checking for integer division by zero. Mips: -- Changed generic-= and generic-/= to not assume that EQ implies =, because it doesn't in the case of NaNs. -- Major rewrite of many things. Merged mumble-immediate SCs into the immediate SC. Wrote several vops to use :constant arg types better. Rewrote all the type testing stuff. New pseudo-atomic, allocators now inline. -- Fixed UNBIND-TO-HERE to not dereference past the end of the binding stack. -- Changed to load the function from static-function-offset relative to NIL instead of computing the symbol and then loading the function. -- Added Miles' change to use JALI instead of LI/JR now that it exists. -- Fixed a lifetime bug in full call. This only showed up if there were either a variable number or > 6 arguments and the caller was large enough that the compute-lra-from-code couldn't be done in one instruction. Byte compilation: Byte compilation reduces the size of the Lisp process by allowing rarely used functions to be compiled more compactly. In comparison with 16f, the basic system image is about 20% smaller. The full system with Hemlock CLX, etc., is also 19% smaller even though it contains new code (the Motif interface) because a larger fraction of the extra systems are byte compiled. Byte compilation overview: The decision to byte-compile or native compile can be done on a per-file or per-code-object basis. COMPILE-FILE now has a :BYTE-COMPILE argument. If T, we byte-compile everything and create a machine-independent fasl file (dependent only on byte order, file type "bytef" or "lbytef".) If :MAYBE (the default, from EXT:*BYTE-COMPILE-DEFAULT*), things are byte-compiled when speed = 0 and debug <= 1. Top-level forms are byte-compiled by default (controlled by ext:*byte-compile-top-level*.) Byte compilation is roughly twice as fast native compilation. Byte compiled code runs 50x--200x slower than native code, and is 6x more dense. This is about 10x faster than the IR1 interpreter, which is itself considered fast in comparison to other Common Lisp interpreters. Historical perspective: this is about as fast as Spice Lisp on a PERQ. Tuning: -- Tweaked GENSYM so that the CONCATENTATE is open-coded. -- Tweaked WITH-ARRAY-DATA to test for (not (array-header-p )) instead of (simple-array * (*)), since that's much faster. This helps all functions that accept non-simple arrays. -- Transform MEMBER, DELETE and ASSOC to MEMQ, ASSQ and DELQ when possible. Inline expand vector & list POSITION, vector FILL and list DELETE-IF. -- Add some INDEX declarations in simple-string concatenate/replace transforms. -- made DIGIT-CHAR-P and DIGIT-WEIGHT maybe-inline. -- Added declarations from efficiency notes in fd-stream, load, package, reader, char and hash. -- Revived the support for FAST-READ-CHAR and the STREAM-IN-BUFFER, which allows READ-CHAR and READ-BYTE to be done with 0 function calls rather than 2. -- Because of above two changes, both the reader and the fasloader are now significantly faster (reader 2x.) -- Default (non-frozen) structure type tests are now significantly faster (no function calls), and somewhat smaller. This and the reader improvement have sped up the compiler somewhat. -- Many debug-info and compiler data structures are now annotated as pure, alloing them to be put in read-only space. This reduces the amound of stuff in static space, speeding GC. -- Real-valued hash-table parameters (rehash-threshold etc.) are canonicalized to single-floats. -- Replaced ISQRT with a much faster version off the net. -- serve-event now uses UNIX-FAST-SELECT, so it can can handle >32 file descriptors and is more efficient. -- Changed UNIX-FAST-SELECT to a macro so that it can be efficient. Changed FD-SET stuff to be efficiently compilable. -- Use an auxiliary function to make the condition for macro arg count errors to save space in macro definitions. -- Byte compile the expander functions for all macros except those in the cold load. Byte-compile various user-interface stuff like the debugger, disassemble and structure print functions. Byte-compile most Hemlock commands. -- Compile PCL's guts unsafe when #+SMALL. -- Some gratuitous RANDOM tuning. Random double floats are now much, much faster. Added transforms for RANDOM to type-specific functions (which can then be inline-expanded). DEFSTRUCT and classes: The structure representation has been changed to point directly to a type descriptor rather than to the symbol type name. This allows faster type tests and better GC support. Also, structure redefinition is now much more conservative; formerly, many cases where code was compiled using differing versions of the same structure were quietly ignored or resulted in strange behavior. Raw allocation of typed slots dramatically increase the efficiency of float-valued slots. Much of DEFSTRUCT has been rewritten, and is now believed to be ANSI complaint. ANSI changes: -- Default defstruct keyword constructors can no longer reference argument values in slot init forms. BOA constructors can still do this, so defining a BOA constructor with all keyword args will have the old effect. -- Class objects are now implemented, see FIND-CLASS, CLASS-NAME, TYPEP, CLASS-OF. TYPE-OF is now based on CLASS-OF, and returns slightly different results than before. -- STRUCTURE-CLASSes now exist. See also the STRUCTURE-OBJECT type. -- BUILT-IN-CLASSes also exist. In some cases CLASS-OF (legally) returns non-standard subclasses of the standard class, e.g. for a float vector, the result is KERNEL::SIMPLE-ARRAY-SINGLE-FLOAT. STANDARD-CHAR and KEYWORD are now DEFTYPEs. Bug fix: -- Typed structures now have the correct (though rather odd) semantics of :offset and :named when inclusion is done. Raw slots: -- Structure slots of subtypes of SINGLE-FLOAT, DOUBLE-FLOAT and (UNSIGNED-BYTE 32) are now allocated in non-descriptor storage, and can be read/written without number-consing overhead. Type tests: -- The default (non-frozen) structure type predicate is now significantly faster (no function call) in the case where the argument is a structure of another type or the type is a supertype of the object's type. The code is also somewhat smaller. Structure Redefinition: -- Handing of structure redefinition is now much more comprehensive. -- Definitions are only considered incompatible when slots have moved or been added, slot types are changed to a type that is not a subtype of the old type, or the inheritance structure has changed. Previously any change at all would produce a warning. -- When a change is incompatible, the default restart (CONTINUE) invalidates old instances, constructors and predicate uses. When speed <= safety, a LAYOUT-INVALID error will be signalled when obsolete instances are passed to a type test (e.g. for type checking.) Use of old code on new instances or old instances when speed > safety > 0 will result in type errors. Other restarts allow you to ignore the redefinition or to clobber the existing information, preserving the old code (in case the change is actually compatible.) -- If the structure length or inheritance structure has changed, an error is signalled when you load code that was compiled with a different structure definition than the one currently in effect. Structure Internals: -- VM:STRUCTURE-USAGE renamed to VM:INSTANCE-USAGE. Internally, the structure type and accessors have also been renamed, e.g. STRUCTURE-REF => %INSTANCE-REF. -- The non-standard STRUCTURE type has become has become EXT:INSTANCE. To (portably) test whether something is really a structure object, do (TYPEP X 'STRUCTURE-OBJECT) PCL: The largest changes are: -- PCL port revamped to re-integrate PCL classes with the type system and to more efficiently dispatch on structure and built-in types. Some miscellaneous tuning. CLOS symbols are now exported from the LISP package, so you don't need to USE-PACKAGE PCL anymore. PCL notes: The integration of PCL with the CMU CL type system has been substantially improved. There are significant improvements in the speed of PCL generic function dispatch of built-in and structure classes and TYPEP of PCL classes. There should also be reduced run-time compilation (e.g. in the Motif inspector) due to less use of non-precompiled dispatch functions. The compiler now recognizes the PCL::CLASS declaration. This clues the compiler in on PCL's knowledge of types (due to being a specialized argument to a method.) CLOS class names are recognized as "real" types by the compiler, not SATISFIES DEFTYPES. Note that LISP:CLASS is still not a PCL class, so PCL needs to shadow CLASS and STANDARD-CLASS. Fixed the #< print function to flame out if *PRINT-READABLY* is true. Supply missing :INITIAL-ELEMENT NIL in MAKE-ARRAY call which can result in a "0 is not a list" or "segment violation" error on redefining a class. September-16-92-PCL-e has been incorporated into the CMU CL sources thanks to Rick Harris. Merge fix to pessimization of GFs with many standard class specializers but also some EQL or built-in class specializers. Graphical debugger/Motif toolkit: We have implemented a new interface to Motif which is functionally similar to CLM, but works better in CMU CL. See: doc/motif-toolkit.doc doc/motif-internals.doc This motif interface has been used to write a new inspector and graphical debugger. There is also a Lisp control panel with a simple file management facility, apropos and inspector dialogs, and controls for setting global options. Call INTERFACE:LISP-CONTROL-PANEL to create the control panel. When INTERFACE:*INTERFACE-STYLE* is :GRAPHICS (the default) and the DISPLAY environment variable is defined, the graphical inspector and debugger will be invoked by INSPECT or when an error is signalled. Possible values are :GRAPHICS and :TTY. If the value is :GRAPHICS, but there is no X display, then we quietly use the TTY interface. Debugger: The command-line debugger now implements the breakpoint and step commands described in the manual: LIST-LOCATIONS [{function | :c}] list the locations for breakpoints. Specify :c for the current frame. Abbreviation: LL LIST-BREAKPOINTS list the active breakpoints. Abbreviations: LB, LBP DELETE-BREAKPOINT [n] remove breakpoint n or all breakpoints. Abbreviations: DEL, DBP BREAKPOINT {n | :end | :start} [:break form] [:function function] [{:print form}*] [:condition form] set a breakpoint. Abbreviations: BR, BP STEP [n] step to the next location or step n times. These commands provide a degree of support for stepping and setting breakpoints inside compiled functions. The variables DEBUG:*USE-BLOCK-STARTS-ONLY* and DEBUG:*PRINT-CODE-LOCATION-KIND* control the verbosity of LIST-LOCATIONS. Enhancements: -- Changed PRINT-FRAME-CALL to print the source if verbosity >= 2 and the source is available. -- Changed source location printing to cache information so that it is much faster when many locations in the same function are printed. The source file is now only printed when the file changes from one printing to the next. Bug fixes: -- Added explicit error checking to the debugger so that we don't get an internal error (bus error in unsafe code, etc.) when attempting to display source from a file that has been excessively modified. -- Bind *BREAK-ON-SIGNALS* to NIL when we call BREAK in SIGNAL so that the debugger doesn't recurse on itself. -- Changed HANDLE-BREAKPOINT in the debugger to allow breakpoints that nobody wants. This can happen if a function-end breakpoint was deactivated while the function was on the stack, because there is no way to convert the bogus-lra back into the real lra. -- Fixed COMPUTE-CALLING-FRAME to not try using %CODE-DEBUG-INFO on things that arn't code components. -- Instead of doing after breakpoints in Lisp, use the new C function breakpoint_do_displaced_inst. That way the C code can do different things on different machines (like use single stepping if available). Misc changes: CLX: We are now distributing version R5.01 of the CLX X library. Among other changes, this is supposed to support the cookie-based host authentication used by OpenWindows. Now compiled with the :CLX-ANSI-COMMON-LISP feature. Pathnames: -- Logical pathnames are now implemented and wildcard matching has been generalized to make them useful. Any namestring beginning with HOST: (where HOST is a defined logical host) will be parsed as a logical pathname. If the prefix is not a defined logical host, it will be still be parsed as a CMU CL "search list." -- Pattern-matching on :WILD-INFERIORS is now implemented, but the Unix code treats it pretty much like :WILD. -- More error cases are detected in TRANSLATE-PATHNAME. -- See new user's manual for CMU CL specific documentation. Stream internals: -- Deleted read-line methods. For simplicity, this rather unimportant operation is now implemented using read-char. -- READ-N-BYTES eof-error-p nil is now mostly non-blocking (it only reads what is in the buffer, or what unix-read returns if the buffer is empty.) To be sure it won't block, you must guard it with a LISTEN. Extensions: -- Changed the stream arg to DISASSEMBLE to be a keyword to avoid optional & keyword lossage. -- Changed PROFILE:PROFILE argument count determination to parse the function type and look at it, instead of trying to fake it. Among other things, this allows efficient profiling of functions with FTYPE declarations even when compilation policy has caused the function-object's type to be dropped. -- Add :CALLERS option to PROFILE which records the most common callers of each profiled function. -- Add new "CPROFILE" package which does instruction-counting profiling with the compiler's assistance. -- Fixed DI:FUNCTION-DEBUG-FUNCTION to work on closures. -- Added EXT:DO-HASH. -- User-defined hashtable tests are now supported. There is a function DEFINE-HASH-TABLE-TEST that takes three arguments: the symbol name of the hash table test, the test function, and the hash function. It updates *hash-table-tests*, which is now internal. The test function takes two objects and returns true iff they are the same. The hash function takes one object and returns two values: the (positive fixnum) hash value and true if the hashing depends on pointer values and will have to be redone if the object moves. -- Added weak hash-table support. - Removed (setf weak-pointer-value) and made make-weak-pointer itself the compiler primitive in order to simplify the gengc port. - Added stuff to fake scavenger hooks in the non-gengc system. -- Moved the RUN-PROGRAM fork/exec stuff into C, for ease in porting. Hemlock: -- In DELETE-BREAKPOINTS-BUFFER-HOOK, if no wire (server died), then don't do anything. -- Changed BEEP flashing to use SLEEP 0.1 instead of DISPLAY-FINISH-OUTPUT because this was causing recursive entry of CLX. -- Fixed from debugger edit command "can't edit source" message not spuriously reinvoke the debugger when invoked from the command line. -- Fix initialization of print-representation attribute so that characters >127 don't cause text to mysteriously disappear. -- Flush carefully-add-font-path call. If people want library:fonts/ in their font-path, they can put it there. This should eliminate the Hemlock initialization error which would happen when using X remotely. Changed SAVE-LISP to no longer save the stacks. Instead, when the core is restored, a (supplied) initial function is invoked which can do whatever kind of setup it wants. This makes a saved lisp totally independent of the location of the C stack, and eliminates the "environment too big" error that happened in some SUNOS environments. A consequence of this is that calling SAVE-LISP terminates the currently running Lisp. SunOS/SPARC: Changed software-version to use /usr/bin/uname instead of stringing the kernel. Removed the load of bit-bash, because we don't want to have to support the assemble routine versions. Significant revamping of startup code (lisp now, not ldb.) The new startup code has better breakpoint support and improved portability. Added :CMU17 to the features list so that PCL can tell if it is in a version 16 or a version 17 series core. Release notes for CMU Common Lisp 16f, 11 December 92 The changes between 16e and 16f are almost exclusively bug-fixes. When we announce a version 17 beta, 16f will probably become the default release (replacing 15e). The PCL has been upgraded from "March 92 (2a)" to "March 92 (2c)", which provides some bug-fixes; also, all of the patches in the March-92-PCL-bugs file have been applied. Enhancements: -- PROVIDE & REQUIRE are now back in the system, since proposed ANSI CL has reinstated them as deprecated features. See the doc strings for these functions and EXT:DEFMODULE. -- The SPARC dynamic heap size limit has been doubled from 64meg to 128meg. Pretty printer: -- Fixed a bug in pprint-let that caused to to barf on (let (nil) ...). -- Fixed pprint-lambda-list to print a space before the dot when the tail of the lambda list is shared. In other words, print (foo . #1=(bar baz)) instead of (foo. #1=(bar baz)). -- Added an additional use of ~^ in pprint-flet so that (flet (nil) ...) doesn't flame out. -- Make pretty printer properly process pprint tabs when output is forced. -- In format pprint logical blocks, fixed ~^ to act like PPRINT-EXIT-IF-LIST-EXHAUSTED instead of blowing out to some containing directive. Compiler: -- Fixed compiler internal error with dead-code deletion of top-level code. -- Bind *gensym-counter* instead of setting it so that compiling doesn't globally reset the gensym counter. -- Fixed problem with interpreted LOAD-TIME-VALUE causing undefined function VALUE-CELL-REF errors. -- Preserve the arglist in interpreted functions for DESCRIBE, etc. -- Gag bound-but-not-referenced warnings when the EXT:INHIBIT-WARNINGS optimize quality is 3. -- Fixed a problem with register allocation conflict analysis which appeared when a function was called with ~>= 50 arguments. -- Fixed bug with optimization of tail local calls. -- Fixed interpreted PROCLAIM/DECLAIM to ignore START-BLOCK, END-BLOCK and declared DECLARATION declarations. Changed the unrecognized proclamation error to be a warning. -- Fixed float heap allocation to not wedge when the store causes a trap (SPARC only.) Trace: -- Added pretty-printer directives so that arg lists and results print better. -- Fixed UNTRACE not to flame out when untracing untraced functions. -- Fixed bug with redefining function traced with encapsulation (e.g. interpreted functions. Misc bug fixes: -- Fixed I/O timeout handling (e.g. for CLX) to correctly borrow from the timeout seconds when computing the new value for the timeout microseconds. -- Restored proper (prompt) handling of queued CLX events in SERVE-EVENT -- Alien enums always take up an int to be compatable with C. Also, sort the from-alist so that enum aliens are unparsed in a canonical format. -- When doing macro destructuring, check to see if some part of a lambda-list is a LIST before checking to see if it is a SYMBOL, because we want NIL to act like the empty list, and not an attempt to bind NIL. -- Fixed PACKAGE-ERROR to have a PACKAGE slot instead of a PATHNAME slot. -- Changed DOLIST not to introduce the spurious let around the result form when there is no result form. Also, just read the var in the spurious let, instead of using IGNORABLE, since the var might be special. -- In complex DEFSETF, don't bother creating temp vars for constants. This is necessary so that keywords stay keywords, and are not changed to gensyms. Enhancements: -- Changed the backquote expanded functions (backq-list, ...) from being inline functions to compiler-macros, since although the optimizer does eventually get the right code, it has to work awful hard. Changed BREAK to accept a condition as well as a format string. -- Exported EXT:CONNECT-TO-UNIX-SOCKET and EXT:CREATE-UNIX-SOCKET, which had been forgotten before. Added code to CONNECT-TO-UNIX-SOCKET so that Unix domain sockets are available for connecting to other processes. -- Removed sys:*task-data* and sys:*task-notify* because they aren't used even under Mach. Release notes for CMU Common Lisp 16e, 5 August 92 16e is primarily a bug-fix release. The main changes from 16d are: -- CLOS support is from March 92 PCL (2a). This is a new version of PCL developed by Richard Harris which incorporates many bug-fixes and ANSI compliance cleanups. He has also back-merged the CMU changes into his sources so that we can release future PCLs without time-consuming merging. On the downside, there are a couple of new bugs (discrimination on pcl::structure-object doesn't always work; generic functions which contain methods which discriminate on both null and list sometimes do not work). Patches for these bugs are available in March-92-PCL-bugs in the CMU CL release area and by anonymous ftp from host parcftp.xerox.com (13.1.64.94), in the directory pub/pcl/. -- TRACE has been reimplemented, has a new syntax and new features. -- The hardcopy and info documentation has been updated. Note that it describes some debugger capabilities (breakpoints) which won't appear until version 17. The fasl file format is the same as for 16d, but some code may need to be recompiled. In particular, the expansion of PPRINT-LOGICAL-BLOCK has changed. March 92 PCL highlights: (see notes.text in the sources for details) -- This version of PCL is much closer than previous versions of PCL to the metaobject protocol specified in "The Art of the Metaobject Protocol", chapters 5 and 6, by Gregor Kiczales, Jim des Riveres, and Daniel G. Bobrow. -- You can use structure-class as a metaclass to create new classes. Classes created this way create and evaluate defstruct forms which have generated symbols for the structure accessors and constructor. -- Various optimization of instance variable access, both inside and outside of methods. -- More work (lookups and precompilation) is done at compile and load time, rather than run-time. New TRACE: Trace has been substantially rewritten, and has a new syntax as well as new functionality: -- Tracing of compiled functions is now implemented using breakpoints. Breakpoints destructively modify the code object, causing all calls to the function to be trapped, instead of only those calls that indirect through the symbol. This makes TRACE more useful for debugging programs that use data structures containing function values, since you can now trace anonymous functions and macros. Also, the breakpoint stops the function after the arguments have been parsed, so arguments can accessed by name in the debugger or in TRACE options. -- Depending on the ENCAPSULATE option and DEBUG:*TRACE-ENCAPSULATE-DEFAULT*, encapsulation may be used instead. This is the default for closures, generic functions and interpreted functions. -- TRACE options are no longer set off by extra parens, and you can specify global trace options which affect all functions traced by a particular call to TRACE. -- Conditional breakpoints now work much better than before. -- *DEBUG-PRINT-LEVEL*, -LENGTH* are used instead of a separate *TRACE-PRINT-LEVEL*, etc. Here is the documentation string (see also the hardcopy/info documentation): TRACE {Option Global-Value}* {Name {Option Value}*}* TRACE is a debugging tool that prints information when specified functions are called. In its simplest form: (trace Name-1 Name-2 ...) TRACE causes a printout on *TRACE-OUTPUT* each time that one of the named functions is entered or returns (the Names are not evaluated.) The output is indented according to the number of pending traced calls, and this trace depth is printed at the beginning of each line of output. Options allow modification of the default behavior. Each option is a pair of an option keyword and a value form. Options may be interspersed with function names. Options only affect tracing of the function whose name they appear immediately after. Global options are specified before the first name, and affect all functions traced by a given use of TRACE. The following options are defined: :CONDITION Form :CONDITION-AFTER Form :CONDITION-ALL Form If :CONDITION is specified, then TRACE does nothing unless Form evaluates to true at the time of the call. :CONDITION-AFTER is similar, but suppresses the initial printout, and is tested when the function returns. :CONDITION-ALL tries both before and after. :WHEREIN Names If specified, Names is a function name or list of names. TRACE does nothing unless a call to one of those functions encloses the call to this function (i.e. it would appear in a backtrace.) Anonymous functions have string names like "DEFUN FOO". :BREAK Form :BREAK-AFTER Form :BREAK-ALL Form If specified, and Form evaluates to true, then the debugger is invoked at the start of the function, at the end of the function, or both, according to the respective option. :PRINT Form :PRINT-AFTER Form :PRINT-ALL Form In addition to the usual prinout, he result of evaluating Form is printed at the start of the function, at the end of the function, or both, according to the respective option. Multiple print options cause multiple values to be printed. :FUNCTION Function-Form This is a not really an option, but rather another way of specifying what function to trace. The Function-Form is evaluated immediately, and the resulting function is traced. :ENCAPSULATE {:DEFAULT | T | NIL} If T, the tracing is done via encapsulation (redefining the function name) rather than by modifying the function. :DEFAULT is the default, and means to use encapsulation for interpreted functions and funcallable instances, breakpoints otherwise. When encapsulation is used, forms are *not* evaluated in the function's lexical environment, but DEBUG:ARG can still be used. :CONDITION, :BREAK and :PRINT forms are evaluated in the lexical environment of the called function; DEBUG:VAR and DEBUG:ARG can be used. The -AFTER and -ALL forms are evaluated in the null environment. Assorted bug fixes and enhancements: System code: -- Changed default base file name for LOAD-FOREIGN to be argv[0] rather than being hard-wired to "lisp". -- Fixed a bad declaration which caused garbage collection to fail if more than MOST-POSITIVE-FIXNUM bytes had been consed since process creation. -- Changed GET-INTERNAL-RUN-TIME to use UNIX-FAST-GETRUSAGE to avoid number-consing and generic arithmetic. Also, rearranged the computation so that the time is correctly computed for up to 457 days, instead of only 71 minutes. -- Merged Miles' fix to MAKE-PATHNAME so that it knows the difference between an arg being NIL and being unsupplied. -- Some partial fixes to circular printing (the #1=#1# bug). PPRINT-LOGICAL-BLOCK no longer checks the list argument for CAR circularity, now that OUTPUT-OBJECT does it for us. -- Fixed reader dispatch macro characters to be case-insensitive, and to disallow digits as sub-characters. -- Changed #A reader to allow arbitrary sequences instead of just lists. -- RUN-PROGRAM now gives a proper error message when "fork" fails (i.e. too many processes.) -- Fixed a bug in initialization of saved cores which caused the old environment to be left on the end of EXT:*ENVIRONMENT-LIST*. One symptom was that RUN-PROGRAM would run programs with strange environment values based on those in effect at the time the core was saved. In particular, Lisp subprocesses (i.e. Hemlock slaves) might get the wrong value of CMUCLLIB, which caused the slave to die before connecting. -- SYSTEM:SERVE-EVENT (and XLIB:EVENT-CASE, etc.) now correctly handle non-integer timeouts. Added declarations to improve the efficiency of event handling. -- Fixed some bugs in UNIX-SELECT which could cause Lisp to hang when more than 32 files were open. Also, improved efficiency in this case. -- Merged Olssons fix to WITH-ENABLED-INTERRUPTS to not try to change interrupt characters anymore. -- A number of bug-fixes for breakpoint support in compiled code (but there are still problems with arbitrary breakpoints.) -- Fixed DI:FRAME-CATCHES CLX: -- Fixed the implementation-dependent pixarray copying routines (for GET-IMAGE, etc.) so that they don't occasionally trash memory, and are actually faster. -- Fixed the definition of the ANGLE type (used by DRAW-ARC, etc.) to work regardless of the kind of real number (single or double float, rational, etc.) -- Fixed several places in image operations where values that could really be negative were declared to be non-negative. Compiler: -- Fixed a bug which caused an internal error whenever a call to random was compiled and the argument type wasn't known to be either a float or an integer. -- Fixed a bug which caused an internal compiler error when a value that wasn't used had an unproven type assertion. -- Fixed some more dead-code deletion bugs. -- Fixed a problem with the new "assignment" optimization of local function call where the compiler could get assertion failures such as tail-sets not being equal. -- Fixed a few places where reoptimization wasn't being triggered when it should have been. -- You can now have a TAGBODY with more than one tag that is non-locally exited to. Evidently this never worked... -- Some changes in debug-info format related to breakpoint support. Misc: -- Fixed some Hemlock Dired commands to know that PATHNAME-DIRECTORY is now a list, not a vector. -- Fixed the bin/sample-wrapper script to use "$@" instead of $* so that arguments are properly passed through. Release notes for CMU Common Lisp 16d, 30 May 92 16d is our first version 16 general release, and incorporates many changes not present in the 15 series. It is currently fairly close to our current internal development (alpha) systems, and is thus less stable. The major changes are: New Aliens New pathnames New pretty printer New format R5.0 CLX. 5/1/90 May Day PCL (REV 4b) Revised manual The fasl file format is nominally compatible with version 15, but the pathname change affects any pathname constants in fasl files, which includes the defined-from information present in every fasl file. It is probably a good idea to recompile everything. CLX and Hemlock are now optional. When CMU CL is installed, the maintainer can elect not to load CLX and Hemlock -- this saves 7 megabytes of disk and improves memory usage somewhat. See the installation section of the README file for details. The ``CMU Common Lisp User's Manual'' has been updated to be more helpful for non-CMU users. The new manual also documents the new Alien facility for foreign function calls and data structure access. The manual is now formatted with Mike Clarkson's LaTeXinfo package, so a consistent version of the documentation is available online in Gnu info format. See `doc/cmu-user.ps' and `doc/cmu-user.info'. General system code: ANSI cleanups: -- ANSI Compiler macros are now implemented: see COMPILER-MACRO-FUNCTION, COMPILER-MACROEXPAND, COMPILER-MACROEXPAND-1 and DEFINE-COMPILER-MACRO. -- Fixed things that invoke *MACROEXPAND-HOOK* to coerce it to a function before calling it. -- Fixed MACRO-FUNCTION to take an environment argument. -- SYMBOL-MACROLET now accepts declarations. -- SHADOW now accepts strings in addition to symbols. -- Added UPGRADED-ARRAY-ELEMENT-TYPE and UPGRADED-COMPLEX-PART-TYPE. -- IGNORABLE is now in the LISP package instead of the EXT package. -- ADJUST-ARRAY has been updated to allow adjusting of arrays which were not created with :adjustable non-nil to be adjusted to new dimensions. -- ADJUSTABLE-ARRAY-P has been updated correspondingly. It returns T if adjust ADJUST-ARRAY would return an EQ array. -- The BASE-CHARACTER type has been renamed to BASE-CHAR. -- The REAL type and REALP function are now implemented. -- Changed the default structure printer to print slot names as keywords instead of unqualified symbols. Enhancements: -- Added MAYBE-INLINE declaration for GET, PUT, etc., so that these functions can be inline expanded according to the compilation policy. -- Added some type declarations so that GET-INTERNAL-REAL-TIME doesn't cons. -- Process the command line before printing the herald so that we can eval some form and quit without printing anything. -- SET now protects against setting T, NIL, and keywords. (SETF SYMBOL-FUNCTION) now expands into FSET, which protects against defining NIL. -- Substantially rearranged function describing to make it more consistent, and added support for describing interpreted functions. -- PURIFY is now called multiple times during system building to improve locality. -- (EVAL-WHEN (EVAL) ...) is now actually eval'ed. Bug fixes: -- Fixed bug in NTH-VALUE where it expanded into bogus code unless ``n'' was a constant integer. -- Fixed FMAKUNBOUND to return the symbol instead of T. -- Allocate memory as executable so that the OS knows to maintain cache consistency. -- Changed DESCRIBE to allow T or NIL as the stream argument. Load enhancements and cleanups: -- The initial value of *LOAD-VERBOSE* is now T. Additionally, LOAD no longer always binds *LOAD-VERBOSE* and *LOAD-PRINT*. Now it only binds them when :VERBOSE or :PRINT are explicity supplied. Therefore, you can set either of these in your init file and it will take effect. -- Normally when *LOAD-VERBOSE* is T, only the file name is printed. Formerly, the loaded stream was always printed, whereas now a stream is printed only when the stream is not a file stream. -- Added ANSI features *LOAD-TRUENAME*, *LOAD-PATHNAME* and *LOAD-PRINT*. -- As per ANSI, bind *READTABLE* to itself to make assignments file-local. -- Added new variables EXT:*SOURCE-FILE-TYPES* and EXT:*OBJECT-FILE-TYPES*. When no file type is specified, LOAD tries the types in these lists to locate the source and object files. LOAD now recognizes source types "l", "cl" and "lsp" in addition to "lisp". -- The compiler OPTIMIZE policy is now bound during load, so proclamations in a file don't leave the global policy clobbered when the load is finished. -- Changed the :IF-SOURCE-NEWER option to signal an error and use restarts, rather than PROMPT-FOR-Y-OR-N. Fixed the load source case to actually load the source, rather than loading the object as a source file... -- Changed load to deal with source files having NIL type more reasonably. -- If a wild pathname is given to LOAD, all files matched will be loaded. -- Proceeding from nonexistent file errors has been improved. It is no longer assumed that missing files are always source files. Added condition restarts for missing files. -- Improved formatting of error and warning messages. Garbage collection: -- Changed the minimal ROOM output to include all easily computed information including whether GC is disabled. The verbose ROOM now conses less. -- Removed the :ENABLE-GC SAVE-LISP option, as it's no longer needed. Garbage collection is now correctly enabled in cores which have been saved and then restarted. -- Added EXT:BYTES-CONSED-BETWEEN-GCS, a function that returns (and sets when used with setf) *BYTES-CONSED-BETWEEN-GCS*. Additionally, it changes *GC-TRIGGER* immediately to reflect the new values of *bytes-consed...*. -- TIME now displays the GC run-time. -- Added EXT:*GC-RUN-TIME* with accumulates the INTERNAL-RUN-TIME spent doing garbage collection. Added declarations to make EXT:GET-BYTES-CONSED more efficient. -- The top-level REP loop now zeros the unused non-zero portion of the control stack to discourage spurious garbage retention. -- The garbage collector now closes open file streams when it reclaims them. Packages: -- The LISP and USER packages have been renamed to COMMON-LISP and COMMON-LISP-USER. LISP and USER are nicknames, so they can still be used. -- The LISP package namespace has been cleaned up somewhat. For example, *DESCRIBE-PRINT-LEVEL* is no longer exported from LISP. -- The Mach/Unix division in the package system has been clarified a great deal. Unix specific features have been moved from the MACH package to the UNIX package. Mach specific features have been left in (or moved to) the MACH package. For example, all standard Unix syscalls are related definitions are un UNIX, whereas vm_statistics remains in MACH and GR-CALL has been moved to MACH. SETF cleanups: -- Changed GET-SETF-METHOD-MULTIPLE-VALUE to try to macroexpand-1 the form when it's an atom in case it's a symbol-macro as per the X3J13 cleanup SYMBOL-MACROLET-SEMANTICS:SPECIAL-FORM. Now you can safely INCF, etc. symbol macros where the macroexpansion has side effects. -- Fixed SETF of GETF to evaluate the various parts in the correct order as per X3J13 cleanup SETF-SUB-METHODS:DELAYED-ACCESS-STORES. -- X3J13 cleanup SETF-MULTIPLE-STORE-VARIABLES: Extend the semantics of the macros SETF, PSETF, SHIFTF, ROTATEF, and ASSERT to allow "places" whose SETF methods have more than one "store variable". In such cases, the macros accept as many values from the newvalue form as there are store variables. As usual, extra values are ignored and missing values default to NIL. -- Extended the long form of DEFSETF to allow the specification of more than one "store variable", with the obvious semantics. -- GET-SETF-METHOD signals an error if there would be more than one store-variable. Printer: Almost all of the printing code has been rewritten/restructured to support all of the printing features added to the language. Some highlights: -- *PRINT-READABLY* is now supported. -- *PRINT-CIRCLE* works irrespective of *PRINT-PRETTY*. Note: the default structure printer currently does not work when *PRINT-CIRCLE* is true: you get #1=#1#. -- *PRINT-LEVEL* abbreviation now works automatically inside structure printers. -- XP has been replaced with a native pretty printer that is fully integrated with the rest of the system. This Supports the ANSI pretty-printing interface instead of the old Waters XP interface. Existing uses of the old interface will need to be updated to use the new names. -- The pretty-printer now unparses backquote forms on printing. To retain this information, the backqoute read macro no longer expands into LIST, CONS, etc. Internal functions are used instead. -- The macros WITH-STANDARD-IO-SYNTAX and PRINT-UNREADABLE-OBJECT have been added. -- All new format. Supports the FORMATTER macro and all the pretty-printing directives. FORMAT has extended to accept a function as the format control (as an alternative to a string.) -- Added support for READTABLE-CASE in symbol printing. Printing when *PRINT-CASE* is :CAPITALIZE and *PRINT-ESCAPE* is NIL is now slightly different than before. Added some missing array type declarations in symbol printing. Bug fixes: -- Fixed a bug which caused some float printing format directives to infinitely loop when a fixed-width field overflowed. -- Specify stream when printing unbound marker. -- Fixed FORMAT to override printer control variables when printing float exponents so that they are always printed in decimal, etc. Reader: ANSI Cleanups: -- *READ-EVAL* is now supported. If a #. is encountered while *READ-EVAL* is NIL (default T), an error is signaled. This intended to allow ``secure'' READ-based command interfaces to be written. -- READTABLE-CASE is now supported. -- The reader now signals the correct type of error when things go wrong instead of always signaling a simple-error. -- Changed the return value of SET-SYNTAX-FROM-CHAR from NIL to T as per X3J13 cleanup RETURN-VALUES-UNSPECIFIED:SPECIFY. [Hard to believe nobody has complained about not conforming to this one.] Bug fixes: -- Fixes to several bugs with respect to #+, #-. In particular, stacked conditionals like "#+foo #-bar baz" now work -- #n= and #n# now detect more error conditions and work on structures. -- # is now a non-terminating macro character, so foo#bar is read as a symbol. -- Added Ted's changes to make INTERNAL-READ-EXTENDED-TOKEN work when there are `|' escapes. The main significance of this is that #+nil '|foo;bar| and #:|foobar| now work properly. Also changed this function to recognize unquoted colons so that #:foo:bar will error, but not #:foo\:bar. Enhancement: -- Export new variable *ignore-extra-close-parentheses* if true (the default), extra close parens are only a warning, not an error. Previously unmatched close parens were quietly ignored. Pathnames: This release supports all the new CltL2 pathname features except for logical pathnames. Following is an overview of the new pathname support. Programs that actually conform to the CLtL1 pathname spec will have very few problems. However, the CLtL1 spec was extremely vague, and CMU CL did not make use of much of the allowed flexibility, so many technically non-portable programs previously worked under CMU CL. The main incompatible changes from CLtL1 to CLtL2: -- Symbols are no longer acceptable filenames. -- PATHNAME-HOST may be any object. -- :UNSPECIFIC is now a legal pathname component. -- MERGE-PATHNAMES now recognizes relative pathnames and treats them specially. The format of directories is now specified (to be a list in a certain format.) This required an incompatible change from the previous practice of using a vector PATHNAME-DIRECTORY and using "DEFAULT" or :ABSOLUTE in the PATHNAME-DEVICE to indicate relative and absolute pathnames. In a related change, the CMU SEARCH-LIST extension was changed so that the search-list now appears in the PATHNAME-DIRECTORY as: (:ABSOLUTE # ...) Other changes to search-lists: -- (SETF SEARCH-LIST) now accepts a string or pathname, and converts it into a one-element list. -- Search-list elements are now canonicalized to pathnames rather than to strings. -- Instead of returning NIL, SEARCH-LIST now signals an error when it is called on an undefined search list. -- SEARCH-LIST-DEFINED-P is a predicate that tells if the search list is currently defined. New features which are now supported: -- Wildcard pathnames are now fully supported. In addition to allowing :WILD in any pathname component, "extended wildcards" such as "foo*.*" are also supported. A consequence of this is that PATTERN objects may appear in wildcard pathname components instead of strings. See PATHNAME-MATCH-P and TRANSLATE-PATHNAME. -- As a CMU CL extension, a wildcard pathname may be used as the argument to any filesystem interface (like OPEN) as long as it matches only one file. -- The pathname :COMMON case mechanism provides a way around the problems of portably specifying string pathname components in the presence of operating systems with differing preferred case in their filesystem. An uppercase string "LISP" is mapped to the "customary case" (lowercase on unix.) Lowercase is also inverted: "readme" becomes "README". Mixed case is left alone. Note that this mechanism is explicitly enabled by supplying :CASE :COMMON to functions such as MAKE-PATHNAME. The default is the old behavior (:CASE :LOCAL). Also, DIRECTORY now actually returns the TRUENAME of each file (as it was always supposed to do.) If a matched file is a symbolic link, the truename may be a file in some other directory that doesn't even match the pattern. The old behavior can be obtained by specifying :FOLLOW-LINKS NIL. The new wildcard pathname mechanism has not yet been used to replace the old single-wildcard matching in Hemlock DIRED, etc. Debugger: -- Added Miles' changes to keep errors and warnings on one line if they fit. -- The debugger now starts up with the error frame as the current frame, so it is no longer necessary to manually skip over internal frames resulting from the error system invocation. -- Fixed some debugger bugs that appeared when debugging interpreted code. -- Added ``DESCRIBE'' debugger command. -- Merged Miles' changes that allow the use of restart names as debugger commands. -- Changed SHOW-RESTARTS to also display the restart name (but only if it's not shadowed by a higher priority restart). Changed the restart command to look for restarts by name if a symbol is typed. -- Bind *CURRENT-LEVEL* to 0, *PRINT-READABLY* to nil, and *READ-EVAL* to T when entering the debugger to make sure things print as expected. The debugger programmer (DEBUG-INTERNALS) interface is now documented in the User's Manual. This interface allows the coding of debuggers and debugger extensions without requiring an intimate understanding of the compiler and run-time system. Be warned that DI features (such as breakpoints) not used by the current debugger may not work very well (wait for version 17.) Debug internals changes: -- DI:DEBUG-FUNCTION-FUNCTION is now implemented. -- Added DI:FLUSH-FRAMES-ABOVE for cleaning up frames to be bound to DEBUG:*STACK-TOP-HINT*. Defstruct: Various fixes and enhancements to defstruct slot parsing and inclusion. It now works to define structures such as: (defstruct super a) (defstruct (sub1 (:conc-name super)) one) (defstruct (sub2 (:conc-name super)) two) (super-a (make-sub1)) previously, a definition such as for SUB2 would define SUPER-A to be a SUB2-specific function, which could then no longer be used on other types. The spec doesn't clearly say that such a construct is legal, but its use seems fairly common. Also, slot parsing is now more rigorous. Unrecognized slot options cause an error, as does a slot spec like: (defstruct foo (a :type t)) Fasl dumping of constant structures has been fixed to conform to X3J13. The main significance of this is that DEFSTRUCT structures are no longer dumpable by default. However, the generic function MAKE-LOAD-FORM isn't really used. Instead, a new defstruct option, :MAKE-LOAD-FORM-FUN, has been added that can be used to specify a function that acts like a MAKE-LOAD-FORM method. When we have a CLOS that supports STRUCTURE-CLASS, the default method for MAKE-LOAD-FORM will use this information instead of having the compiler use it directly. The old behavior can be enabled on a structure by structure basis by using the :MAKE-LOAD-FORM-FUN defstruct option: (defstruct (foo (:make-load-form-fun :just-dump-it-normally)) ...) Compiler: Cleanups: -- Added the LOAD-TIME-VALUE support special form. -- Displaced or adjustable arrays and vectors with fill pointers can now be dumped in fasl files, but are converted to simple array with the same elements. -- Arrays of floats are left as arrays of floats instead of being converted into arrays of element-type T that just so happen to hold a bunch of floats. -- Changed IGNORE and IGNORABLE to recognize #'fn-name for declaring that local functions are not used. Exported IGNORABLE from LISP. New optimizations: -- Iterations written using tail recursion are now optimized through a special-casing of local functions where all calls but one are tail-recursive self-calls. Such functions are compiled with no environment manipulation, resulting in the same code as explicit iteration. -- Loop rotation (Knuth "while" loop optimization) been added. This is the optimization that negates the loop exit test and places it at the end of the loop, and then jumps there at loop entry. Note that this is part of control optimization, and not simply a recoding of certain iteration macros. In fact, for historical reasons DO, etc. already had the exit test negated, but the compiler was cleverly un-negating the test. -- Flow analysis now recognizes function calls and special forms which do not yield any value because they unwind or signal an error. This results in improved code for error checks, since the compiler can get by with fewer unconditional branches. A function can be declared not to return by declaring its result type to be NIL (not to be confused with NULL). If a function declared NIL does return, an error will be signalled. -- Added derive-type methods for ASIN, ACOS, ACOSH, ATANH and SQRT which figure out whether the result type is real on the basis of the argument range. Also, the result of ATAN is always real, so we don't need a result type assertion. -- Added optimization which deletes MULTIPLE-VALUE-BINDs when all variables have been deleted. -- Eliminated some spurious checking of the result types of safe VOPs (such as the SPARC tadd for fixnum arithmetic.) -- Transform uses of the SEARCH generic sequence function on simple strings into a call to a more efficient string-specific function. -- Added multiplier recoding of (UNSIGNED-BYTE 32) multiplication. This converts 32 bit (or smaller) unsigned multiplication by any compile-time constant into a shift-add sequence. This is much faster when the constant is a near power of two or when general multiplication is slow (as on the SPARC.) -- Added comprehensive handling of arithmetic and logical identities when an arg is -1, 0 or +1. -- Added a RANDOM derive-type method: (random 13) is (integer 0 12). Enhancements: -- Changed the compiler to temporarily increase *BYTES-CONSED- BETWEEN-GCS* by a factor of 4 during compilation of each top-level form, instead of turning off all garbage collection. -- Bind *PRINT-LINES* around compiler error output to *ERROR-PRINT-LINES*. -- Do not warn about undefined variables, functions or types when the INHIBIT-WARNINGS OPTIMIZE quality is 3. -- Some optimizations are now considered "important"; failure of important optimizations causes an efficiency note even when speed=inhibit-warnings (i.e. by default.) -- Print a error summary even when *COMPILE-VERBOSE* is false. (This is only printed when there are errors, so this doesn't seem a violation of the spirit of the spec.) Bug fixes: -- Merged Miles' fix to disassembly of the MIPS coprocessor move instructions. -- Fixed spelling of "efficency" in exported variables such as *EFFICIENCY-NOTE-COST-THRESHOLD* -- Fixed some cases where incomplete optimization could happen. -- Fixed some arithmetic "identities" that failed to preserve the sign of -0.0. -- Fixed TYPES-INTERSECT to consider any supertype of T to definitely intersect with anything (even an unknown type.) -- Fixed a problem where inconsistent declarations could fail to be detected at compile time. -- Fixed the VALUES transform (which discards unused subforms) to work on (VALUES). -- More bug fixes to dead code deletion. -- Fixed a problem where a special binding would not be removed on exit from the scope if there was no code in the scope (e.g. it had all been deleted.) -- Fixed handling of named (DEFCONSTANT) constants so that EQ relationships are properly preserved. Extensions: -- Export FEATUREP from EXT. This takes a feature expression and tests it against the value of *FEATURES*. Allow LISP:AND, LISP:OR, and LISP:NOT in features lists in addition to :AND, :OR, and :NOT. This makes featurep useful outside of #+ and #-. -- The encapsulation mechanism (similar to advise facilities) has been revamped to work on SETF function names as well as symbols. EXT:ENCAPSULATED-DEFINITION Returns whatever definition is stored for name, regardless of whether it is encapsulated. Unlike FDEFINITION, this does not strip off the encapsulation. This is SETF'able. EXT:ENCAPSULATE Replaces the definition of name with a function that binds name's arguments a variable named argument-list, binds name's definition to a variable named basic-definition, and EVAL's body in that context. Type is whatever you would like to associate with this encapsulation for identification in case you need multiple encapsuations of the same name. EXT:UNENCAPSULATE Removes name's most recent encapsulation of the specified type. EXT:ENCAPSULATED-P Returns t if name has an encapsulation of the given type, otherwise nil. EXT:*SETF-FDEFINITION-HOOK* A list of functions invoked by (SETF FDEFINITION) before storing the new value. Each hook function must take the function name and the new-value. Hemlock: -- Fixed the :FILE branch of "Help on Parse" to trim leading directory components off the pathname if it wouldn't otherwise fit on the screen. -- Fixed GET-EDITOR-TTY-INPUT to not assume that UNIX-READ will always work. -- When we reallocate string table vectors to grow them, clear the old vector so that it won't hold onto garbage (in case the vector was in static space, but pointed to dynamic values.) This was a major cause of memory leakage in Hemlock. -- Added a (setf (ts-stream-char-pos stream) 0) to the accept-input function so that char-pos will be reset to zero whenever the user presses enter. PCL: -- The version has been updated to "5/1/90 May Day PCL (REV 4b)". -- The Code walker now understands the real SYMBOL-MACROLET, and the PCL macro definition is no longer used. -- Fixed a bug in WALK-ARGLIST where it would ignore the rest of the current arglist if the current arg destructured. This was causing it to compile forms like: (macrolet ((foo ((a b) c) ...)) ...) as: (macrolet ((foo ((a b)) ...)) ...) note the loss of the arg c. System: -- All the SAP-REF- functions now take byte offsets. Previously, the 16 and 32bit versions were scaled by the element size. -- Fixed UNIX-IOCTL to not flame out of the cmd is a ub-32 instead of a sb-32. -- Added Miles' TCSETPGRP, TCGETPGRP, and TTY-PROCESS-GROUP. -- Unix syscalls are now more restrictive in the kind of arguments that they accept. In particular, UNIX:UNIX-READ, etc., no longer automatically accept a vector (or string) argument. You must use VECTOR-SAP to convert the vector to a system area pointer. Note that WITHOUT-GCING should be wrapped around any syscall which is passed a pointer to the Lisp heap, since the object might otherwise move doing the syscall. -- Changed LOAD-FOREIGN to be exported from ALIEN. Changed it have keyword args instead of optionals. Deleted obsolete linker argument.