# SBCL Manual ###### [in package SB-MANUAL] This is an unofficial rendering of the SBCL Manual using MGL-PAX with the same content as the official version at but with heavy linking internally, to the clhs, and to the source code on [GitHub](https://github.com/sbcl/sbcl). The output is for SBCL version 2.6.7.109-ed0f3c5, generated 2026-08-13 03:20:30. See for this document in other formats. This manual is part of the SBCL software system. See the README file for more information. The manual is largely derived from the manual for the [CMUCL](https://cmucl.org/) system, which was produced at Carnegie Mellon University and later released into the public domain. The manual is in the public domain and is provided with absolutely no warranty. See the COPYING and CREDITS files for more information. ## Getting Support and Reporting Bugs ### Volunteer Support Your primary source of SBCL support should probably be the mailing list sbcl-help: in addition to other users SBCL developers monitor this list and are available for advice. As an anti-spam measure subscription is required for posting: Remember that the people answering your question are volunteers, so you stand a much better chance of getting a good answer if you ask a good question. Before sending mail, check the list archives at either or to see if your question has been answered already. Checking the bug database is also worth it (see Reporting Bugs), to see if the issue is already known. For general advice on asking good questions, see . ### Commercial Support There is no formal organization developing SBCL, but if you need a paid support arrangement or custom SBCL development, we maintain the list of companies and consultants below. Use it to identify service providers with appropriate skills and interests, and contact them directly. The SBCL project cannot verify the accuracy of the information or the competence of the people listed, and they have provided their own blurbs below: you must make your own judgement of suitability from the available information - refer to the links they provide, the CREDITS file, mailing list archives, CVS commit messages, and so on. Please feel free to ask for advice on the sbcl-help list. (At present, no companies or consultants wish to advertise paid support or custom SBCL development in this manual). ### Reporting Bugs SBCL uses Launchpad to track bugs. The bug database is available at Reporting bugs there requires registering at Launchpad. However, bugs can also be reported on the mailing list sbcl-bugs, which is moderated but does not require subscribing. Simply send email to [sbcl-bugs@lists.sourceforge.net](mailto:sbcl-bugs@lists.sourceforge.net) and the bug will be checked and added to Launchpad by SBCL maintainers. See the HACKING file on how to send patches. #### How to Report Bugs Effectively Please include enough information in a bug report that someone reading it can reproduce the problem, i.e. don't write Subject: apparent bug in PRINT-OBJECT (or *PRINT-LENGTH*?) PRINT-OBJECT doesn't seem to work with *PRINT-LENGTH*. Is this a bug? but instead Subject: apparent bug in PRINT-OBJECT (or *PRINT-LENGTH*?) In sbcl-1.2.3 running under OpenBSD 4.5 on my Alpha box, when I compile and load the file (DEFSTRUCT (FOO (:PRINT-OBJECT (LAMBDA (X Y) (LET ((*PRINT-LENGTH* 4)) (PRINT X Y))))) X Y) then at the command line type (MAKE-FOO) the program loops endlessly instead of printing the object. A more in-depth discussion on reporting bugs effectively can be found at . #### How to Report Signal-related Bugs If you run into a signal related bug, you are getting fatal errors such as signal N is [un]blocked or just hangs, and you want to send a useful bug report then: - Compile SBCL with ldb enabled (feature :sb-ldb, see base-target-features.lisp-expr). - Isolate a smallish test case, run it. - If it just hangs kill it with sigabrt: kill -ABRT . - Print the backtrace from ldb by typing ba. - Attach gdb: gdb -p and get backtraces for all threads: thread apply all ba. - If multiple threads are in play then still in gdb, try to get Lisp backtrace for all threads: thread apply all call backtrace_from_fp($ebp, 100, 0). Substitute $ebp with $rbp on x86-64. The backtraces will appear in the stdout of the SBCL process. - Send a report with the backtraces and the output (both stdout and stderr) produced by SBCL. - Don't forget to include OS and SBCL version. - If available, include information on outcome of the same test with other versions of SBCL, OS, ... ## Introduction SBCL is a mostly-conforming implementation of the ANSI Common Lisp standard. This manual focuses on behavior which is specific to SBCL, not on behavior which is common to all implementations of ANSI Common Lisp. ### ANSI Conformance Essentially every type of non-conformance is considered a bug. (The exceptions involve internal inconsistencies in the standard.) See Reporting Bugs. - prog2 returns the primary value of its second form, as specified in the Arguments and Values section of the specification for that operator, not that of its first form, as specified in the Description. - The string type is considered to be the union of all types (array c (size)) for all non-nil subtypes c of character, excluding arrays specialized to the empty type. - The :order long form option in define-method-combination method group specifiers accepts the value nil as well as :most-specific-first and :most-specific-last, in order to allow programmers to declare that the order of methods playing that role in the method combination does not matter. ### Extensions SBCL comes with numerous extensions, some in core and some in modules loadable with require. Unfortunately, not all of these extensions have proper documentation yet. - System Definition Tool: ASDF is a flexible and popular protocol-oriented system definition tool by Daniel Barlow. - Foreign Function Interface: The sb-alien package allows interfacing with C-code, loading shared object files, etc. See Foreign Function Interface. sb-grovel can be used to partially automate generation of foreign function interface definitions. - Recursive Event Loop: SBCL provides a recursive event loop (serve-event) for doing non-blocking IO on multiple streams without using threads. - Timeouts and Deadlines: SBCL allows restricting the execution time of individual operations or parts of a computation using :timeout arguments to certain blocking operations, synchronous timeouts and asynchronous timeouts. The latter two affect operations without explicit timeout support (such as standard functions and macros). See Timeouts and Deadlines. - Metaobject Protocol: The sb-mop package provides an implementation of the metaobject protocol for the Common Lisp Object System as described in The Art of the Metaobject Protocol by Kiczales et al. - Extensible Sequences: SBCL allows users to define subclasses of the sequence class. See Extensible Sequences. - Native Threads: SBCL has native threads on numerous platforms, capable of taking advantage of SMP on multiprocessor machines. See Threading. - Network Interface: The sb-bsd-sockets module is a low-level networking interface, providing both TCP and UDP sockets. See Networking. - Introspective Facilities: The sb-introspect module offers numerous introspective extensions, including access to function lambda-lists and a cross referencing facility. - Operating System Interface: The sb-ext package contains a number of functions for running external processes, accessing environment variables, etc. The sb-posix module provides a lispy interface to standard POSIX facilities. - Extensible Streams: The package sb-gray provides an implementation of Gray Streams. The Simple Streams module is an implementation of the Simple Streams API proposed by Franz Inc. - Profiling: The sb-profile package provides an exact, per-function Deterministic Profiler. The sb-sprof module is SBCL's Statistical Profiler, capable of call-graph generation and instruction level profiling, which also supports allocation profiling. - Customization Hooks: SBCL contains a number of extra-standard customization hooks that can be used to tweak the behaviour of the system. See Customization Hooks for Users. - sb-aclrepl: The sb-aclrepl module provides an Allegro-style toplevel for SBCL, as an alternative to the classic CMUCL-style one. - CLTL2 Compatibility Layer: The SB-CLTL2 module provides sb-cltl2:compiler-let and environment access functionality described in Common Lisp The Language, 2nd Edition which were removed from the language during the ANSI standardization process. - Executable Delivery: The :executable argument to sb-ext:save-lisp-and-die can produce a "standalone" executable containing both an image of the current Lisp session and an SBCL runtime. - Bitwise Rotation: The sb-rotate-byte module provides an efficient primitive for bitwise rotation of integers, an operation required by e.g. numerous cryptographic algorithms but not available as a primitive in ANSI Common Lisp. - Test Harness: The sb-rt module is a simple yet attractive regression and unit-test framework. - MD5 Sums: The sb-md5 module provides an implementation of the MD5 message digest algorithm for Common Lisp, using the Modular Arithmetic optimizations provided by SBCL. ### Idiosyncrasies The information in this section describes some of the ways that SBCL deals with choices that the ANSI standard leaves to the implementation. #### Declarations Declarations are generally treated as assertions. This general principle, and its implications, and the bugs which still keep the compiler from quite satisfying this principle, are discussed in Declarations as Assertions. #### FASL format SBCL fasl-format is binary compatible only with the exact SBCL version it was generated with. While this is obviously suboptimal, it has proven more robust than trying to maintain fasl compatibility across versions: accidentally breaking things is far too easy, and can lead to hard to diagnose bugs. The following snippet handles fasl recompilation automatically for ASDF-based systems, and makes a good candidate for inclusion in the user or system initialization file (see Initialization Files). (require :asdf) ;;; If a fasl was stale, try to recompile and load (once). (defmethod asdf:perform :around ((o asdf:load-op) (c asdf:cl-source-file)) (handler-case (call-next-method o c) ;; If a fasl was stale, try to recompile and load (once). (sb-ext:invalid-fasl () (asdf:perform (make-instance 'asdf:compile-op) c) (call-next-method)))) #### Compiler-only Implementation SBCL is essentially a compiler-only implementation of Common Lisp. That is, for all but a few special cases, eval creates a lambda expression, calls compile on the lambda expression to create a compiled function, and then calls funcall on the resulting function object. A more traditional interpreter is also available on default builds; it is usually only called internally. This is explicitly allowed by the ANSI standard but leads to some oddities; e.g. at default settings, functionp and compiled-function-p are equivalent, and they collapse into the same function when SBCL is built without the interpreter. #### Defining Constants SBCL is quite strict about ANSI's definition of defconstant. ANSI says that doing defconstant of the same symbol more than once is undefined unless the new value is eql to the old value. Conforming to this specification is a nuisance when the "constant" value is only constant under some weaker test like string= or equal. It's especially annoying because, in SBCL, defconstant takes effect not only at load time but also at compile time, so that just compiling and loading reasonable code like (defconstant +foobyte+ '(1 4)) runs into this undefined behavior. Many implementations of Common Lisp try to help the programmer around this annoyance by silently accepting the undefined code and trying to do what the programmer probably meant. SBCL instead treats the undefined behavior as an error. Often such code can be rewritten in portable ANSI Common Lisp which has the desired behavior. E.g., the code above can be given an exactly defined meaning by replacing defconstant either with defparameter or with a customized macro which does the right thing, e.g. (defmacro define-constant (name value &optional doc) `(defconstant ,name (if (boundp ',name) (symbol-value ',name) ,value) ,@(when doc (list doc)))) or possibly along the lines of the sb-int:defconstant-eqx macro used internally in the implementation of SBCL itself. In circumstances where this is not appropriate, the programmer can handle the condition type sb-ext:defconstant-uneql and choose either the continue restart or abort restart as appropriate. #### Style Warnings SBCL gives style warnings about various kinds of perfectly legal code, e.g. - multiple defuns of the same symbol in different units; - special variables not named in the conventional *foo* style, and lexical variables unconventionally named in the *foo* style. This causes friction with people who point out that other ways of organizing code (especially avoiding the use of defgeneric) are just as aesthetically stylish. However, these warnings should be read not as warning, bad aesthetics detected, you have no style but as warning, this style keeps the compiler from understanding the code as well as you might like. That is, unless the compiler warns about such conditions, there's no way for the compiler to warn about some programming errors which would otherwise be easy to overlook. (Related bug: The warning about multiple defuns is pointlessly annoying when you compile and then load a function containing defun wrapped in eval-when, and ideally should be suppressed in that case, but still isn't as of SBCL 0.7.6.) ### Development Tools #### Editor Integration Though SBCL can be used running "bare", the recommended mode of development is with an editor connected to SBCL, supporting not only basic lisp editing (paren-matching, etc), but providing among other features an integrated debugger, interactive compilation, and automated documentation lookup. Currently SLIME (Superior Lisp Interaction Mode for Emacs) together with Emacs is recommended for use with SBCL, though other options exist as well. Historically, the ILISP package at provided similar functionality, but it does not support modern SBCL versions. SLIME can be downloaded from . #### Language Reference CLHS (Common Lisp HyperSpec) is a hypertext version of the ANSI standard, made freely available by LispWorks -- an invaluable reference. See . #### Generating Executables SBCL can generate stand-alone executables. The generated executables include the SBCL runtime itself, so no restrictions are placed on program functionality. For example, a deployed program can call compile and load, which requires the compiler to be present in the executable. For further information, sb-ext:save-lisp-and-die. ### More SBCL Information #### SBCL Homepage The SBCL website at has some general information, plus links to mailing lists devoted to SBCL, and to archives of these mailing lists. Subscribing to the mailing lists sbcl-help and sbcl-announce is recommended: both are fairly low-volume, and help you keep abreast with SBCL development. #### Online Documentation Documentation for non-ANSI extensions for various commands is available online from the SBCL executable itself. The extensions for functions which have their own command prompts (e.g. the debugger, and inspect) are documented in text available by typing help at their command prompts. The extensions for functions which don't have their own command prompt (such as trace) are described in their documentation strings, unless your SBCL was compiled with an option not to include documentation strings, in which case the documentation strings are only readable in the source code. #### Additional Documentation Files Besides this user manual both SBCL source and binary distributions include some other SBCL-specific documentation files, which should be installed along with this manual on your system, e.g. in /usr/local/share/doc/sbcl/. - copying: Licence and copyright summary. - credits: Authorship information on various parts of SBCL. - install: Covers installing SBCL from both source and binary distributions on your system, and also has some installation related troubleshooting information. - news: Summarizes changes between various SBCL versions. #### Internals Documentation If you're interested in the development of the SBCL system itself, then subscribing to [sbcl-devel@lists.sourceforge.net](mailto:sbcl-devel@lists.sourceforge.net) is a good idea. SBCL internals documentation -- besides comments in the source -- is available in the Web Archive: . Some low-level information describing the programming details of the conversion from CMUCL to SBCL is available in the doc/FOR-CMUCL-DEVELOPERS file. ### More Common Lisp Information #### Internet Community IRC channels on : - #common-lisp: "Common Lisp, the #1=(programmable . #1#) programming language" - #lispcafe: "The Lisp Café; sit down, have a drink, chat about anything, and enjoy your stay. | | Be insuperable to each other". - #sbcl: "Steel Bank Common Lisp Dev Hangout" You can use or a normal IRC client. Also, see , as well as and , which contain numerous pointers places in the net where lispers talks shop. #### Third-party Libraries For a wealth of information about free Common Lisp libraries and tools we recommend checking out CLiki: . The most popular library manager is Quicklisp: . #### Common Lisp Books If you're not a programmer and you're trying to learn, many introductory Lisp books are available. However, we don't have any standout favorites. If you are an experienced programmer in other languages but need to learn about Common Lisp, some books stand out: - Practical Common Lisp, by Peter Seibel An excellent introduction to the language, covering both the basics and "advanced topics" like macros, CLOS, and packages. Available both in print format and on the web: . - Paradigms Of Artificial Intelligence Programming, by Peter Norvig Good information on general Common Lisp programming, and many nontrivial examples. Whether or not your work is AI, it's a very good book to look at. - On Lisp, by Paul Graham An in-depth treatment of macros, but not recommended as a first Common Lisp book, since it is slightly pre-ANSI so you need to be on your guard against non-standard usages, and since it doesn't really even try to cover the language as a whole, focusing solely on macros. Downloadable from . - Object-Oriented Programming In Common Lisp, by Sonya Keene With the exception of Practical Common Lisp, most introductory books don't emphasize CLOS. This one does. Even if you're very knowledgeable about object oriented programming in the abstract, it's worth looking at this book if you want to do any OO in Common Lisp. Some abstractions in CLOS (especially multiple dispatch) go beyond anything you'll see in most OO systems, and there are a number of lesser differences as well. This book tends to help with the culture shock. - Art Of Metaobject Programming, by Gregor Kiczales et al. Currently the prime source of information on the Common Lisp Metaobject Protocol, which is supported by SBCL. Section 2 (Chapters 5 and 6) are freely available at . ### History and Implementation of SBCL You can work productively with SBCL without knowing or understanding anything about where it came from, how it is implemented, or how it extends the ANSI Common Lisp standard. However, a little knowledge can be helpful in order to understand error messages, to troubleshoot problems, to understand why some parts of the system are better debugged than others, and to anticipate which known bugs, known performance problems, and missing extensions are likely to be fixed, tuned, or added. SBCL is descended from CMUCL, which is itself descended from Spice Lisp, including early implementations for the Mach operating system on the IBM RT, back in the 1980s. Some design decisions from that time are still reflected in the current implementation: - The system expects to be loaded into a fixed-at-compile-time location in virtual memory, and also expects the location of all of its heap storage to be specified at compile time. - The system overcommits memory, allocating large amounts of address space from the system (often more than the amount of virtual memory available) and then failing if it ends up using too much of the allocated storage. - The system is implemented as a C program which is responsible for supplying low-level services and loading a Lisp .core file. SBCL also inherited some newer architectural features from CMUCL. The most important is that on some architectures it has a generational GC, which has various implications (mostly good) for performance. These are discussed in another chapter, Efficiency. SBCL has diverged from CMUCL in that SBCL is now essentially a compiler-only implementation of Common Lisp. This is a change in implementation strategy, taking advantage of the freedom "any of these facilities might share the same execution strategy" guaranteed in clhs 3.1 (Evaluation). It does not mean SBCL can't be used interactively, and in fact the change is largely invisible to the casual user, since SBCL still can and does execute code interactively by compiling it on the fly. (It is visible if you know how to look, like using compiled-function-p; and it is visible in the way that SBCL doesn't have many bugs which behave differently in interpreted code than in compiled code.) What it means is that in SBCL, the eval function only truly "interprets" a few easy kinds of forms, such as symbols which are boundp. More complicated forms are evaluated by calling compile and then calling funcall on the returned result. The direct ancestor of SBCL is the x86 port of CMUCL. This port was in some ways the most cobbled-together of all the CMUCL ports, since a number of strange changes had to be made to support the register-poor x86 architecture. Some things (like tracing and debugging) do not work particularly well there. SBCL should be able to improve in these areas (and has already improved in some other areas), but it takes a while. On the x86 SBCL -- like the x86 port of CMUCL -- uses a conservative GC. This means that it doesn't maintain a strict separation between tagged and untagged data, instead treating some untagged data (e.g. raw floating point numbers) as possibly-tagged data and so not collecting any Lisp objects that they point to. This has some negative consequences for average time efficiency (though possibly no worse than the negative consequences of trying to implement an exact GC on a processor architecture as register-poor as the X86) and also has potentially unlimited consequences for worst-case memory efficiency. In practice, conservative GCs work reasonably well, not getting anywhere near the worst case. But they can occasionally cause odd patterns of memory usage. The fork from CMUCL was based on a major rewrite of the system bootstrap process. CMUCL has for many years tolerated a very unusual "build" procedure which doesn't actually build the complete system from scratch, but instead progressively overwrites parts of a running system with new versions. This quasi-build procedure can cause various bizarre bootstrapping hangups, especially when a major change is made to the system. It also makes the connection between the current source code and the current executable more tenuous than in other software systems -- it's easy to accidentally build a CMUCL system containing characteristics not reflected in the current version of the source code. Other major changes since the fork from CMUCL include: - SBCL has removed many CMUCL extensions, (e.g. IP networking, remote procedure call, Unix system interface, and X11 interface) from the core system. Most of these are available as contributed modules (distributed with SBCL) or third-party modules instead. - SBCL has deleted or deprecated some nonstandard features and code complexity which helped efficiency at the price of maintainability. For example, the SBCL compiler no longer implements memory pooling internally (and so is simpler and more maintainable, but generates more garbage and runs more slowly). ## Starting and Stopping ### Starting SBCL #### Running from Shell To run SBCL, type sbcl at the command line. You should end up in the toplevel REPL (read-eval-print loop), where you can interact with SBCL by typing expressions. $ sbcl This is SBCL 0.8.13.60, an implementation of ANSI Common Lisp. More information about SBCL is available at . SBCL is free software, provided as is, with absolutely no warranty. It is mostly in the public domain; some portions are provided under BSD-style licenses. See the CREDITS and COPYING files in the distribution for more information. * (+ 2 2) 4 * (exit) $ Also see Command Line Options and Stopping SBCL. #### Running from Emacs To run SBCL as an inferior-lisp from Emacs, in your .emacs do something like: ;;; The SBCL binary and command-line arguments (setq inferior-lisp-program "/usr/local/bin/sbcl --noinform") For more information on using SBCL with Emacs, see Editor Integration. #### Shebang Scripts Standard Unix tools that are interpreters follow a common command line protocol that is necessary to work with "shebang scripts". SBCL supports this via the --script command line option (see Command Line Options). Example file (hello.lisp): #!/usr/local/bin/sbcl --script (write-line "Hello, World!") Usage from the command line: $ ./hello.lisp Hello, World! Note that SBCL skips the shebang line when it reads the file: $ sbcl --script hello.lisp Hello, World! ### Stopping SBCL #### Exit SBCL can be stopped at any time by calling sb-ext:exit, optionally returning a specified numeric value to the calling process. See Threading for information about terminating individual threads. - [function] sb-ext:exit &key code abort (timeout *exit-timeout*) Terminates the process, causing SBCL to exit with code. code defaults to 0 when abort is false, and 1 when it is true. When abort is false (the default), current thread is first unwound, *exit-hooks* are run, other threads are terminated, and standard output streams are flushed before SBCL calls exit(3) -- at which point atexit(3) functions will run. If multiple threads call exit with abort being false, the first one to call it will complete the protocol. When abort is true, SBCL exits immediately by calling _exit(2) without unwinding stack, or calling exit hooks. Note that _exit(2) does not call atexit(3) functions unlike exit(3). Recursive calls to exit cause exit to behave as if abort was true. timeout controls waiting for other threads to terminate when abort is nil. Once current thread has been unwound and *exit-hooks* have been run, spawning new threads is prevented and all other threads are terminated by calling sb-thread:terminate-thread on them. The system then waits for them to finish using sb-thread:join-thread, waiting at most a total timeout seconds for all threads to join. Those threads that do not finish in time are simply ignored while the exit protocol continues. timeout defaults to *exit-timeout*, which in turn defaults to 60. timeout nil means to wait indefinitely. Note that timeout applies only to sb-thread:join-thread, not *exit-hooks*. Since sb-thread:terminate-thread is asynchronous, getting multithreaded application termination with complex cleanups right using it can be tricky. To perform an orderly synchronous shutdown use an exit hook instead of relying on implicit thread termination. Consequences are unspecified if serious conditions occur during exit excepting errors from *exit-hooks*, which cause warnings and stop execution of the hook that signaled, but otherwise allow the exit process to continue normally. #### End of File By default SBCL also exits on end of input, caused either by user pressing Control-D on an attached terminal, or end of input when using SBCL as part of a shell pipeline. #### Saving a Core Image SBCL has the ability to save its state as a file for later execution. This functionality is important for its bootstrapping process, and is also provided as an extension to the user. - [function] sb-ext:save-lisp-and-die core-file-name &key (toplevel #'toplevel-init) executable save-runtime-options callable-exports (purify t) root-structures (environment-name "auxiliary") compression Save a "core image", i.e. enough information to restart a Lisp process later in the same state, in the file of the specified name. Only global state is preserved: the stack is unwound in the process. The following &key arguments are defined: - :toplevel The function to run when the created core file is resumed. The default function handles command line toplevel option processing (see Toplevel Options) and runs the top level read-eval-print loop. This function returning is equivalent to (sb-ext:exit :code 0) being called. toplevel functions should always provide an abort restart: otherwise code they call will run without one. - :executable If true, arrange to combine the SBCL runtime and the core image to create a standalone executable. If false (the default), the core image will not be executable on its own. Executable images always behave as if they were passed the --noinform runtime option. If :executable is :elf-object, then the resulting core will be wrapped in a .o which requires further linking. (EXPERIMENTAL) - :save-runtime-options If true, values of runtime options --dynamic-space-size and --control-stack-size that were used to start SBCL are stored in the standalone executable, and restored when the executable is run. This also inhibits normal runtime option processing, causing all command line arguments to be passed to the toplevel. If :accept-runtime-options then --dynamic-space-size and --control-stack-size are still processed by the runtime. Meaningless if :executable is nil. - :callable-exports This should be a list of symbols to be initialized to the appropriate alien callables on startup. All exported symbols should be present as global symbols in the symbol table of the runtime before the saved core is loaded. When this list is non-empty, the :toplevel argument cannot be supplied. - :purify If true (the default), then some objects in the restarted core will be memory-mapped as read-only. Among those objects are numeric vectors that were determined to be compile-time constants, and any immutable values according to the language specification such as symbol names. - :root-structures This should be a list of the main entry points in any newly loaded systems. This need not be supplied, but locality and/or gc performance may be better if they are. This has two different but related meanings: If :purify is true - and only for cheneygc - the root structures are those which anchor the set of objects moved into static space. On gencgc - and only on platforms supporting immobile code - these are the functions and/or function-names which commence a depth-first scan of code when reordering based on the statically observable call chain. The complete set of reachable objects is not affected per se. This argument is meaningless if neither enabling precondition holds. - :environment-name This has no purpose; it is accepted only for legacy compatibility. - :compression This is only meaningful if the runtime was built with the :sb-core-compression feature enabled. If nil (the default), saves to uncompressed core files. If :sb-core-compression was enabled at build-time, the argument may also be an integer from -7 to 22, corresponding to zstd compression levels, or t (which is equivalent to the default compression level, 9). - :application-type Present only on Windows and is meaningful only with :executable t. Specifies the subsystem of the executable, :console or :gui. The notable difference is that :gui doesn't automatically create a console window. The default is :console. The save/load process changes the values of some global variables: - *standard-output*, *debug-io*, etc Everything related to open streams is necessarily changed, since the OS won't let us preserve a stream across save and load. - *default-pathname-defaults* This is reinitialized to reflect the working directory where the saved core is loaded. save-lisp-and-die interacts with sb-alien:load-shared-object: see its documentation for details. On threaded platforms only a single thread may remain running after sb-ext:*save-hooks* have run. Applications using multiple threads can be save-lisp-and-die friendly by registering a save-hook that quits any additional threads, and an init-hook that restarts them. This implementation is not as polished and painless as you might like: - It corrupts the current Lisp image enough that the current process needs to be killed afterwards. This can be worked around by forking another process that saves the core. - There is absolutely no binary compatibility of core images between different runtime support programs. Even runtimes built from the same sources at different times are treated as incompatible for this purpose. This isn't because we like it this way, but just because there don't seem to be good quick fixes for either limitation and no one has been sufficiently motivated to do lengthy fixes. - [variable] sb-ext:*save-hooks* nil A list of function designators which are called in an unspecified order before creating a saved core image. Unused by SBCL itself: reserved for user and applications. In cases where the standard initialization files have already been loaded into the saved core, and alternative ones should be used (or none at all), SBCL allows customizing the initfile pathname computation. - [variable] sb-ext:*sysinit-pathname-function* # Designator for a function of zero arguments called to obtain a pathname designator for the default sysinit file, or nil. If the function returns nil, no sysinit file is used unless one has been specified on the command-line. - [variable] sb-ext:*userinit-pathname-function* # Designator for a function of zero arguments called to obtain a pathname designator or a stream for the default userinit file, or nil. If the function returns nil, no userinit file is used unless one has been specified on the command-line. To facilitate distribution of SBCL applications using external resources, the filesystem location of the SBCL core file being used is available from Lisp. - [variable] sb-ext:*core-pathname* "" The absolute pathname of the running SBCL core. #### Exit on Errors SBCL can also be configured to exit if an unhandled error occurs, which is mainly useful for acting as part of a shell pipeline; doing so under most other circumstances would mean giving up large parts of the flexibility and robustness of Common Lisp. See Debugger Entry and the command line option --disable-debugger in Runtime Options. ### Command Line Options Command line options can be considered an advanced topic; for ordinary interactive use, no command line arguments should be necessary. In order to understand the command line argument syntax for SBCL, it is helpful to understand that the SBCL system is implemented as two components, a low-level runtime environment written in C and a higher-level system written in Common Lisp itself. Some command line arguments are processed during the initialization of the low-level runtime environment, some command line arguments are processed during the initialization of the Common Lisp system, and any remaining command line arguments are made available to user code via sb-ext:*posix-argv*. The full, unambiguous syntax for invoking SBCL at the command line is: sbcl * --end-runtime-options \ * --end-toplevel-options \ * For convenience, --end-runtime-options and --end-toplevel-options can be omitted, which can be convenient when you are running the program interactively, and you can see that no ambiguities are possible with the option values you are using. Omitting these elements is probably a bad idea for any batch file where any of the options are under user control, since it makes it impossible for SBCL to detect erroneous command line input, so that erroneous command line arguments will be passed on to the user program even if they was intended for the runtime system or the Lisp system. #### Runtime Options - --core Run the specified Lisp core file instead of the default. Note that if the Lisp core file is a user-created core file, it may run a nonstandard toplevel which does not recognize the standard toplevel options. - --dynamic-space-size Size of the dynamic space reserved on startup in megabytes. Default value is platform dependent. - --control-stack-size Size of control stack reserved for each thread in megabytes. Default value is 2. - --tls-limit Maximum number of thread-local symbols in threaded builds. Default value is 4096. - --noinform Suppress the printing of any banner or other informational message at startup. This makes it easier to write Lisp programs which work cleanly in Unix pipelines. See also the --noprint and --disable-debugger options. - --disable-ldb Disable the low-level debugger. Only effective if SBCL is compiled with ldb. disabling ldb - --lose-on-corruption There are some dangerous low-level errors (for instance, control stack exhausted, memory fault) that (or whose handlers) can corrupt the image. By default, SBCL prints a warning, then tries to continue and handle the error in Lisp, but this will not always work, and SBCL may malfunction or even hang. With this option, upon encountering such an error, SBCL will exit instead of invoking ldb (if present and enabled enabling ldb). - --script As a runtime option, this is equivalent to --noinform --disable-ldb --lose-on-corruption --end-runtime-options --script . See the description of --script as a toplevel option below. If there are no other command line arguments following --script, the filename argument can be omitted. - --merge-core-pages When platform support is present, provide hints to the operating system that identical pages may be shared between processes until they are written to. This can be useful to reduce the memory usage on systems with multiple SBCL processes started from similar but differently-named core files, or from compressed cores. Without platform support, do nothing. By default only compressed cores trigger hinting. - --no-merge-core-pages Ensures that no sharing hint is provided to the operating system. - --help Print some basic information about SBCL, then exit. - --version Print SBCL's version information, then exit. In the future, runtime options may be added to control behaviour such as lazy allocation of memory. Runtime options, including any --end-runtime-options option, are stripped out of the command line before the Lisp toplevel logic gets a chance to see it. #### Toplevel Options The following options are processed and removed by the default toplevel (see sb-ext:save-lisp-and-die). - --sysinit Load filename instead of the default system initialization file (see Initialization Files). - --no-sysinit Don't load a system-wide initialization file. If this option is given, the --sysinit option is ignored. - --userinit Load filename instead of the default user initialization file (see Initialization Files.) - --no-userinit Don't load a user initialization file. If this option is given, the --userinit option is ignored. - --eval After executing any initialization file, but before starting the read-eval-print loop on standard input, read and evaluate command. More than one --eval option can be used, and all will be read and executed, in the order they appear on the command line. - --load This is equivalent to --eval '(load "")'. The special syntax is intended to reduce quoting headaches when invoking SBCL from shell scripts. - --noprint When ordinarily the toplevel "read-eval-print loop" would be executed, execute a "read-eval loop" instead, i.e. don't print a prompt and don't echo results. Combined with the --noinform runtime option, this makes it easier to write Lisp "scripts" which work cleanly in Unix pipelines. - --disable-debugger By default when SBCL encounters an error, it enters the builtin debugger, allowing interactive diagnosis and possible intercession. This option disables the debugger, causing errors to print a backtrace and exit with status 1 instead. When given, this option takes effect before loading of initialization files or processing --eval and --load options. See sb-ext:disable-debugger and Debugger Entry. - --script Implies --no-userinit --no-sysinit --disable-debugger --end-toplevel-options. Causes the system to load the specified file instead of entering the read-eval-print-loop, and exit afterwards. If the file begins with a shebang line, it is ignored. If there are no other command line arguments following, the filename can be omitted: this causes the script to be loaded from standard input instead. Shebang lines in standard input script are currently not ignored. In either case, if there is an unhandled error (e.g. end of file, or a broken pipe) on either standard input, standard output, or standard error, the script silently exits with code 0. This allows e.g. safely piping output from SBCL to head -n1 or similar. Additionally, the option sets *compile-verbose* and *load-verbose* to nil while loading the file to avoid potentially verbose diagnostic messages printed on the standard output. ### Initialization Files SBCL processes initialization files with read and eval, not load; hence initialization files can be used to set startup *package* and *readtable*, and for proclaiming a global optimization policy. - System Initialization File: Defaults to $SBCL_HOME/sbclrc, or if that doesn't exist to /etc/sbclrc. Can be overridden with the command line option --sysinit or --no-sysinit (see Toplevel Options). The system initialization file is intended for system administrators and software packagers to configure locations of installed third party modules, etc. - User Initialization File: Defaults to $HOME/.sbclrc. Can be overridden with the command line option --userinit or --no-userinit (see Toplevel Options). The user initialization file is intended for personal customizations, such as loading certain modules at startup, defining convenience functions to use in the REPL, handling automatic recompilation of FASLs (see FASL format), etc. Neither initialization file is required. ### Initialization and Exit Hooks SBCL provides hooks into the system initialization and exit. - [variable] sb-ext:*init-hooks* nil A list of function designators which are called in an unspecified order when a saved core image starts up, after the system itself has been initialized, but before non-user threads such as the finalizer thread have been started. Unused by SBCL itself: reserved for user and applications. - [variable] sb-ext:*exit-hooks* nil A list of function designators which are called in an unspecified order when SBCL process exits. Unused by SBCL itself: reserved for user and applications. Using (sb-ext:exit :abort t), or calling exit(3) directly circumvents these hooks. ## Compiler This chapter will discuss most compiler issues other than efficiency, including compiler error messages, the SBCL compiler's unusual approach to type safety in the presence of type declarations, the effects of various compiler optimization policies, and the way that inlining and open coding may cause optimized code to differ from a naive translation. Efficiency issues are sufficiently varied and separate that they have their own chapter, Efficiency. ### Diagnostic Messages #### Controlling Verbosity The compiler can be quite verbose in its diagnostic reporting, rather more then some users would prefer -- the amount of noise emitted can be controlled, however. To control emission of compiler diagnostics (of any severity other than error: Diagnostic Severity) use the sb-ext:muffle-conditions and sb-ext:unmuffle-conditions declarations, specifying the type of condition that is to be muffled (the muffling is done using an associated muffle-warning restart). Global control: ;;; Muffle compiler-notes globally (declaim (sb-ext:muffle-conditions sb-ext:compiler-note)) Local control: ;;; Muffle compiler-notes based on lexical scope (defun foo (x) (declare (optimize speed) (fixnum x) (sb-ext:muffle-conditions sb-ext:compiler-note)) (values (* x 5) ; no compiler note from this (locally (declare (sb-ext:unmuffle-conditions sb-ext:compiler-note)) ;; this one gives a compiler note (* x -5)))) - [declaration] sb-ext:muffle-conditions Syntax: (sb-ext:muffle-conditions &rest types). Muffle the diagnostic messages that would be caused by compile-time signals of types. - [declaration] sb-ext:unmuffle-conditions Syntax: (sb-ext:muffle-conditions &rest types). Cancel the effect of a previous sb-ext:muffle-conditions declaration. Various details of how the compiler messages are printed can be controlled via the alist sb-ext:*compiler-print-variable-alist*. - [variable] sb-ext:*compiler-print-variable-alist* nil An association list describing new bindings for special variables to be used by the compiler for error-reporting, etc. E.g. ((*print-length* . 10) (*print-level* . 6) (*print-pretty* . nil)). The variables in the car positions are bound to the values in the cdr during the execution of some debug commands. When evaluating arbitrary expressions in the debugger, the normal values of the printer control variables are in effect. Initially empty, *compiler-print-variable-alist* is typically used to specify bindings for printer control variables. For information about muffling warnings signaled outside of the compiler, see Customization Hooks for Users. #### Diagnostic Severity There are four levels of compiler diagnostic severity: - error - warning - style warning - note The first three levels correspond to condition classes which are defined in the ANSI standard for Common Lisp and which have special significance to the compile and compile-file functions. These levels of compiler error severity occur when the compiler handles conditions of these classes. The fourth level of compiler error severity, note, corresponds to the sb-ext:compiler-note, and is used for problems which are too mild for the standard condition classes, typically hints about how efficiency might be improved. The sb-ext:code-deletion-note, a subtype of sb-ext:compiler-note, is signalled when the compiler deletes user-supplied code after proving that the code in question is unreachable. Future work for SBCL includes expanding this hierarchy of types to allow more fine-grained control over emission of diagnostic messages. - [condition] sb-ext:compiler-note Root of the hierarchy of conditions representing information discovered by the compiler that the user might wish to know, but which does not merit a style-warning (or any more serious condition). - [condition] sb-ext:code-deletion-note simple-compiler-note A condition type signalled when the compiler deletes code that the user has written, having proved that it is unreachable. #### Understanding Compiler Diagnostics The messages emitted by the compiler contain a lot of detail in a terse format, so they may be confusing at first. The messages will be illustrated using this example program: (defmacro zoq (x) `(roq (ploq (+ ,x 3)))) (defun foo (y) (declare (symbol y)) (zoq y)) The main problem with this program is that it is trying to add 3 to a symbol. Note also that the functions roq and ploq aren't defined anywhere. ##### Parts of a Compiler Diagnostic When processing this program, the compiler will produce this warning: ; file: /tmp/foo.lisp ; in: DEFUN FOO ; (ZOQ Y) ; --> ROQ PLOQ ; ==> ; (+ Y 3) ; ; caught WARNING: ; Asserted type NUMBER conflicts with derived type (VALUES SYMBOL &OPTIONAL). In this example we see each of the six possible parts of a compiler diagnostic: - file: /tmp/foo.lisp is the name of the file that the compiler read the relevant code from. The file name is displayed because it may not be immediately obvious when there is an error during compilation of a large system, especially when with-compilation-unit is used to delay undefined warnings. - in: DEFUN FOO is the definition top level form responsible for the diagnostic. It is obtained by taking the first two elements of the enclosing form whose first element is a symbol beginning with def. If there is no such enclosing def form, then the outermost form is used. If there are multiple def forms, then they are all printed from the outside in, separated by =>s. In this example, the problem was in the defun for foo. - (zoq y) is the original source form responsible for the diagnostic. Original source means that the form directly appeared in the original input to the compiler, i.e. in the lambda passed to compile or in the top level form read from the source file. In this example, the expansion of the zoq macro was responsible for the message. - --> roq ploq This is the processing path that the compiler used to produce the code that caused the message to be emitted. The processing path is a representation of the evaluated forms enclosing the actual source that the compiler encountered when processing the original source. The path is the first element of each form, or the form itself if the form is not a list. These forms result from the expansion of macros or source-to-source transformation done by the compiler. In this example, the enclosing evaluated forms are the calls to roq and ploq. These calls resulted from the expansion of the zoq macro. - ==> (+ y 3) is the actual source responsible for the diagnostic. If the actual source appears in the explanation, then we print the next enclosing evaluated form, instead of printing the actual source twice. (This is the form that would otherwise have been the last form of the processing path.) In this example, the problem is with the evaluation of the reference to the variable y. - caught WARNING: Asserted type NUMBER conflicts with derived type (VALUES SYMBOL &OPTIONAL). is the explanation of the problem. In this example, the problem is that, while the call to + requires that its arguments are all of type number, the compiler has derived that Y will evaluate to a symbol. Note that (values symbol &optional) expresses that y evaluates to precisely one value. Note that each part of the message is distinctively marked: - file: and in: mark the file and definition, respectively. - The original source is an indented form with no prefix. - Each line of the processing path is prefixed with -->. - The actual source form is indented like the original source, but is marked by a preceding ==> line. (FIXME: no it isn't.) - The explanation is prefixed with the diagnostic severity, which can be caught ERROR:, caught WARNING:, caught STYLE-WARNING:, or note:. Each part of the message is more specific than the preceding one. If consecutive messages are for nearby locations, then the front part of the messages would be the same. In this case, the compiler omits as much of the second message as in common with the first. For example: ; file: /tmp/foo.lisp ; in: DEFUN FOO ; (ZOQ Y) ; --> ROQ ; ==> ; (PLOQ (+ Y 3)) ; ; caught STYLE-WARNING: ; undefined function: PLOQ ; ==> ; (ROQ (PLOQ (+ Y 3))) ; ; caught STYLE-WARNING: ; undefined function: ROQ In this example, the file, definition and original source are identical for the two messages, so the compiler omits them in the second message. If consecutive messages are entirely identical, then the compiler prints only the first message, followed by: [Last message occurs times] where is the number of times the message was given. If the source was not from a file, then no file line is printed. If the actual source is the same as the original source, then the processing path and actual source will be omitted. If no forms intervene between the original source and the actual source, then the processing path will also be omitted. ##### Original and Actual Source The original source displayed will almost always be a list. If the actual source for an message is a symbol, the original source will be the immediately enclosing evaluated list form. So even if the offending symbol does appear in the original source, the compiler will print the enclosing list and then print the symbol as the actual source (as though the symbol were introduced by a macro.) When the actual source is displayed (and is not a symbol), it will always be code that resulted from the expansion of a macro or a source-to-source compiler optimization. This is code that did not appear in the original source program; it was introduced by the compiler. Keep in mind that when the compiler displays a source form in an diagnostic message, it always displays the most specific (innermost) responsible form. For example, compiling this function (defun bar (x) (let (a) (declare (fixnum a)) (setq a (foo x)) a)) gives this error message ; file: /tmp/foo.lisp ; in: DEFUN BAR ; (LET (A) ; (DECLARE (FIXNUM A)) ; (SETQ A (FOO X)) ; A) ; ; caught WARNING: ; Asserted type FIXNUM conflicts with derived type (VALUES NULL &OPTIONAL). This message is not saying that there is a problem somewhere in this let -- it is saying that there is a problem with the let itself. In this example, the problem is that a's nil initial value is not a fixnum. ##### Processing Paths The processing path is mainly useful for debugging macros, so if you don't write macros, you can probably ignore it. Consider this example: (defun foo (n) (dotimes (i n *undefined*))) Compiling results in this error message: ; in: DEFUN FOO ; (DOTIMES (I N *UNDEFINED*)) ; --> DO BLOCK LET TAGBODY RETURN-FROM ; ==> ; (PROGN *UNDEFINED*) ; ; caught WARNING: ; undefined variable: *UNDEFINED* Note that do appears in the processing path. This is because dotimes expands into: (do ((i 0 (1+ i)) (#:g1 n)) ((>= i #:g1) *undefined*) (declare (type unsigned-byte i))) The rest of the processing path results from the macroexpansion of do: (block nil (let ((i 0) (#:g1 n)) (declare (type unsigned-byte i)) (tagbody (go #:g3) #:g2 (psetq i (1+ i)) #:g3 (unless (>= i #:g1) (go #:g2)) (return-from nil (progn *undefined*))))) In this example, the compiler descended into the block, let, tagbody and return-from to reach the progn printed as the actual source. This is a place where the "actual source appears in explanation" rule was applied. The innermost actual source form was the symbol undefined itself, but that also appeared in the explanation, so the compiler backed out one level. ### Handling of Types One of the most important features of the SBCL compiler (similar to the original CMUCL compiler, also known as Python) is its fairly sophisticated understanding of the Common Lisp type system and its conservative approach to the implementation of type declarations. These two features reward the use of type declarations throughout development, even when high performance is not a concern. Also, as discussed in the chapter on performance (see Efficiency), the use of appropriate type declarations can be very important for performance as well. The SBCL compiler also has a greater knowledge of the Common Lisp type system than other compilers. Support is incomplete only for types involving the satisfies type specifier. #### Declarations as Assertions The SBCL compiler treats type declarations differently from most other Lisp compilers. Under default compilation policy the compiler doesn't blindly believe type declarations, but considers them assertions about the program that should be checked: all type declarations that have not been proven to always hold are asserted at runtime. Remaining bugs in the compiler's handling of types unfortunately provide some exceptions to this rule, see Implementation Limitations. CLOS slot types form a notable exception. Types declared using the :type slot option in defclass are asserted if and only if the class was defined in safe code and the slot access location is in safe code as well. This laxness does not pose any internal consistency issues, as the CLOS slot types are not available for the type inferencer, nor do CLOS slot types provide any efficiency benefits. There are three type checking policies available in SBCL, selectable via optimize declarations. - Full Type Checks All declarations are considered assertions to be checked at runtime, and all type checks are precise. The default compilation policy provides full type checks. Used when (or (>= safety 2) (>= safety speed 1)). - Weak Type Checks Declared types may be simplified into faster to check supertypes: for example, (or (integer -17 -7) (integer 7 17)) is simplified into (integer -17 17). > Warning: It is relatively easy to corrupt the heap when > weak type checks are used if the program contains type-errors. Used when (and (< safety 2) (< safety speed)). - No Type Checks All declarations are believed without assertions. Also disables argument count and array bounds checking. > Warning: Any type errors in code where type checks are not > performed are liable to corrupt the heap. Used when (= safety 0). #### Precise Type Checking Precise checking means that the check is done as though typep had been called with the exact type specifier that appeared in the declaration. If a variable is declared to be (integer 3 17), then its value must always be an integer between 3 and 17. If multiple type declarations apply to a single variable, then all the declarations must be correct; it is as though all the types were intersected producing a single and type specifier. To gain maximum benefit from the compiler's type checking, you should always declare the types of function arguments and structure slots as precisely as possible. This often involves the use of or, member, and other list-style type specifiers. #### Getting Existing Programs to Run Since SBCL's compiler does much more comprehensive type checking than most Lisp compilers, SBCL may detect type errors in programs that have been debugged using other compilers. These errors are mostly incorrect declarations, although compile-time type errors can find actual bugs if parts of the program have never been tested. Some incorrect declarations can only be detected by run-time type checking. It is very important to initially compile a program with full type checks (high safety optimization) and then test this safe version. After the checking version has been tested, then you can consider weakening or eliminating type checks. This applies even to previously debugged programs because the SBCL compiler does much more type inference than other Common Lisp compilers, so an incorrect declaration can do more damage. The most common problem is with variables whose constant initial value doesn't match the type declaration. Incorrect constant initial values will always be flagged by a compile-time type error, and they are simple to fix once located. Consider this code fragment: (prog (foo) (declare (fixnum foo)) (setq foo ...) ...) Here foo is given an initial value of nil but is declared to be a fixnum. Even if it is never read, the initial value of a variable must match the declared type. There are two ways to fix this problem. Change the declaration (prog (foo) (declare (type (or fixnum null) foo)) (setq foo ...) ...) or change the initial value (prog ((foo 0)) (declare (fixnum foo)) (setq foo ...) ...) It is generally preferable to change to a legal initial value rather than to weaken the declaration, but sometimes it is simpler to weaken the declaration than to try to make an initial value of the appropriate type. Another declaration problem occasionally encountered is incorrect declarations on defmacro arguments. This can happen when a function is converted into a macro. Consider this macro: (defmacro my-1+ (x) (declare (fixnum x)) `(the fixnum (1+ ,x))) Although legal and well-defined Common Lisp code, this meaning of this definition is almost certainly not what the writer intended. For example, this call is illegal: (my-1+ (+ 4 5)) This call is illegal because the argument to the macro is (+ 4 5), which is a list, not a fixnum. Because of macro semantics, it is hardly ever useful to declare the types of macro arguments. If you really want to assert something about the type of the result of evaluating a macro argument, then put a the in the expansion: (defmacro my-1+ (x) `(the fixnum (1+ (the fixnum ,x)))) In this case, it would be stylistically preferable to change this macro back to a function and declare it inline. Some more subtle problems are caused by incorrect declarations that can't be detected at compile time. Consider this code: (do ((pos 0 (position #a string :start (1+ pos)))) ((null pos)) (declare (fixnum pos)) ...) Although pos is almost always a fixnum, it is nil at the end of the loop. If this example is compiled with full type checks (the default), then running it will signal a type error at the end of the loop. If compiled without type checks, the program will go into an infinite loop (or perhaps position will complain because (1+ nil) isn't a sensible start.) Why? Because if you compile without type checks, the compiler just quietly believes the type declaration. Since the compiler believes that pos is always a fixnum, it believes that pos is never nil, so (null pos) is never true, and the loop exit test is optimized away. Such errors are sometimes flagged by unreachable code notes, but it is still important to initially compile and test any system with full type checks, even if the system works fine when compiled using other compilers. In this case, the fix is to weaken the type declaration to (or fixnum null). (Actually, this declaration is unnecessary in SBCL, since it already knows that position returns a non-negative fixnum or nil.) Note that there is usually little performance penalty for weakening a declaration in this way. Any numeric operations in the body can still assume that the variable is a fixnum, since nil is not a legal numeric argument. Another possible fix would be to say: (do ((pos 0 (position #a string :start (1+ pos)))) ((null pos)) (let ((pos pos)) (declare (fixnum pos)) ...)) This would be preferable in some circumstances, since it would allow a non-standard representation to be used for the local pos variable in the loop body. #### Implementation Limitations If an ftype is placed after the function definition the function won't perform any type checks, and the calls to the function will blindly trust the declared types. (optimize (debug 3)) will not trust any ftype declarations. ### Compiler Policy Compiler policy is controlled by the optimize declaration, supporting all ANSI optimization qualities (debug, safety, space, and speed). (A deprecated extension sb-ext:inhibit-warnings is still supported but liable to go away at any time.) For effects of various optimization qualities on type-safety and debuggability see Declarations as Assertions and Debugger Policy Control. Ordinarily, when the speed quality is high, the compiler emits notes to notify the programmer about its inability to apply various optimizations. For selective muffling of these notes, see Controlling Verbosity. The value of space mostly influences the compiler's decision whether to inline operations, which tend to increase the size of programs. Use the value 0 with caution, since it can cause the compiler to inline operations so indiscriminately that the net effect is to slow the program by causing cache misses or even swapping. - [function] sb-ext:describe-compiler-policy &optional spec Print all global optimization settings, augmented by spec. - [function] sb-ext:restrict-compiler-policy &optional quality (min 0) (max 3) Assign a minimum value to an optimization quality. quality is the name of the optimization quality to restrict, min (defaulting to zero) is the minimum allowed value, and max (defaults to 3) is the maximum. Returns the alist describing the current policy restrictions. If quality is nil or not given, nothing is done. Otherwise, if min is zero or max is 3 or neither are given, any existing restrictions of quality are removed. See also :policy option in with-compilation-unit. - [macro] with-compilation-unit options &body body Affects compilations that take place within its dynamic extent. It is intended to be eg. wrapped around the compilation of all files in the same system. Following options are defined: - :override One of the effects of this form is to delay undefined warnings until the end of the form, instead of giving them at the end of each compilation. If override is nil (the default), then the outermost with-compilation-unit form grabs the undefined warnings. Specifying :override true causes that form to grab any enclosed warnings, even if it is enclosed by another with-compilation-unit. - :policy Provides dynamic scoping for global compiler optimization qualities and restrictions, limiting effects of subsequent optimize proclamations and calls to sb-ext:restrict-compiler-policy to the dynamic scope of body. If :override is false, the specified :policy is merged with current global policy. If :override is true, current global policy, including any restrictions, is discarded in favor of the specified :policy. Supplying :policy nil is equivalent to the option not being supplied at all, i.e. dynamic scoping of policy does not take place. This option is an SBCL-specific experimental extension: Interface subject to change. - :source-namestring Attaches the value returned by the to the internal debug-source information as the namestring of the source file. Normally the namestring of the input-file for compile-file is used: this option can be used to provide source-file information for functions compiled using compile, or to override the input-file of compile-file. If both an outer and an inner with-compilation-unit provide a :source-namestring, the inner one takes precedence. Unaffected by :override. This is an SBCL-specific extension. - :source-plist Attaches the value returned by the to internal debug-source information of functions compiled in within the dynamic extent of body. Primarily for use by development environments, in order to eg. associate function definitions with editor-buffers. Can be accessed using sb-introspect:definition-source-plist. If an outer with-compilation-unit form also provide a source-plist, it is appended to the end of the provided source-plist. Unaffected by :override. This is an SBCL-specific extension. Examples: ;; Prevent proclamations from the file leaking, and restrict ;; SAFETY to 3 -- otherwise uses the current global policy. (with-compilation-unit (:policy '(optimize)) (restrict-compiler-policy 'safety 3) (load "foo.lisp")) ;; Using default policy instead of the current global one, ;; except for DEBUG 3. (with-compilation-unit (:policy '(optimize debug) :override t) (load "foo.lisp")) ;; Same as if :POLICY had not been specified at all: SAFETY 3 ;; proclamation leaks out from WITH-COMPILATION-UNIT. (with-compilation-unit (:policy nil) (declaim (optimize safety)) (load "foo.lisp")) Also, see the CLHS. ### Compiler Errors #### Type Errors at Compile Time If the compiler can prove at compile time that some portion of the program cannot be executed without a type error, then it will give a warning at compile time. It is possible that the offending code would never actually be executed at run-time due to some higher level consistency constraint unknown to the compiler, so a type warning doesn't always indicate an incorrect program. For example, consider this code fragment: (defun raz (foo) (let ((x (case foo (:this 13) (:that 9) (:the-other 42)))) (declare (fixnum x)) (foo x))) Compilation produces this warning: ; in: DEFUN RAZ ; (CASE FOO (:THIS 13) (:THAT 9) (:THE-OTHER 42)) ; --> LET COND IF COND IF COND IF ; ==> ; (COND) ; ; caught WARNING: ; This is not a FIXNUM: ; NIL In this case, the warning means that if foo isn't any of :this, :that or :the-other, then x will be initialized to nil, which the fixnum declaration makes illegal. The warning will go away if ecase is used instead of case, or if :the-other is changed to t. This sort of spurious type warning happens moderately often in the expansion of complex macros and in inline functions. In such cases, there may be dead code that is impossible to correctly execute. The compiler can't always prove this code is dead (could never be executed), so it compiles the erroneous code (which will always signal an error if it is executed) and gives a warning. #### Errors During Macroexpansion The compiler handles errors that happen during macroexpansion, turning them into compiler errors. If you want to debug the error (to debug a macro), you can set *break-on-signals* to error. For example, this definition: (defun foo (e l) (do ((current l (cdr current)) ((atom current) nil)) (when (eq (car current) e) (return current)))) gives this error: ; in: DEFUN FOO ; (DO ((CURRENT L (CDR CURRENT)) ; ((ATOM CURRENT) NIL)) ; (WHEN (EQ (CAR CURRENT) E) (RETURN CURRENT))) ; ; caught ERROR: ; (in macroexpansion of (DO # #)) ; (hint: For more precise location, try *BREAK-ON-SIGNALS*.) ; DO step variable is not a symbol: (ATOM CURRENT) #### Read Errors SBCL's compiler does not attempt to recover from read errors when reading a source file, but instead just reports the offending character position and gives up on the entire source file. ### Open Coding and Inline Expansion Since Common Lisp forbids the redefinition of standard functions, the compiler can have special knowledge of these standard functions embedded in it. This special knowledge is used in various ways (open coding, inline expansion, source transformation), but the implications to the user are basically the same: - Attempts to redefine standard functions may be frustrated, since the function may never be called. Although it is technically illegal to redefine standard functions, users sometimes want to implicitly redefine these functions when they are debugging using the trace macro. Special-casing of standard functions can be inhibited using the notinline declaration, but even then some phases of analysis such as type inferencing are applied by the compiler. - The compiler can have multiple alternate implementations of standard functions that implement different trade-offs of speed, space and safety. This selection is based on the Compiler Policy. When a function call is open coded, inline code whose effect is equivalent to the function call is substituted for that function call. When a function call is closed coded, it is usually left as is, although it might be turned into a call to a different function with different arguments. As an example, if nthcdr were to be open coded, then (nthcdr 4 foobar) might turn into (cdr (cdr (cdr (cdr foobar)))) or even (do ((i 0 (1+ i)) (list foobar (cdr foobar))) ((= i 4) list)) If nth is closed coded, then (nth x l) might stay the same, or turn into something like (car (nthcdr x l)) In general, open coding sacrifices space for speed, but some functions (such as car) are so simple that they are always open-coded. Even when not open-coded, a call to a standard function may be transformed into a different function call (as in the last example) or compiled as static call. Static function call uses a more efficient calling convention that forbids redefinition. ### Interpreter By default SBCL implements eval by calling the native code compiler. SBCL also includes an interpreter for use in special cases where using the compiler is undesirable, for example due to compilation overhead. Unlike in some other Lisp implementations, in SBCL interpreted code is not safer or more debuggable than compiled code. - [variable] sb-ext:*evaluator-mode* :compile Toggle between different evaluator implementations. If set to :compile, an implementation of eval that calls the compiler will be used. If set to :interpret, an interpreter will be used. ### Advanced Compiler Use and Efficiency Hints For more advanced usages of the compiler, please see the chapter of the same name in the CMUCL manual. Many aspects of the compiler have stayed exactly the same, and there is a much more detailed explanation of the compiler's behavior and how to maximally optimize code in their manual. In particular, while SBCL no longer supports byte-code compilation, it does support CMUCL's block compilation facility allowing whole program optimization and increased use of the local call convention. Unlike CMUCL, SBCL is able to open-code forward-referenced type tests while block compiling. This helps for mutually referential defstructs in particular. ## Debugger This chapter documents the debugging facilities of SBCL, including the debugger, single-stepper and trace, and the effect of (optimize debug) declarations. ### Debugger Entry #### Debugger Banner When you enter the debugger, it looks something like this: debugger invoked on a TYPE-ERROR in thread 11184: The value 3 is not of type LIST. You can type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL. restarts (invokable by number or by possibly-abbreviated name): 0: [ABORT ] Reduce debugger level (leaving debugger, returning to toplevel). 1: [TOPLEVEL] Restart at toplevel READ/EVAL/PRINT loop. (CAR 1 3) 0] The first group of lines describe what the error was that put us in the debugger. In this case car was called on 3, causing a type-error. This is followed by the "beginner help line", which appears only if sb-debug:*debug-beginner-help-p* is true (default). Next comes a listing of the active restart names, along with their descriptions -- the ways we can restart execution after this error. In this case, both options return to top-level. Restarts can be selected by entering the corresponding number or name. The current frame appears right underneath the restarts, immediately followed by the debugger prompt. #### Debugger Invocation The debugger is invoked when: - error is called, and the condition it signals is not handled. - break is called, or signal is called with a condition that matches the current *break-on-signals*. - The debugger is explicitly entered with the invoke-debugger function. When the debugger is invoked by a condition, ANSI mandates that the value of *debugger-hook*, if any, be called with two arguments: the condition that caused the debugger to be invoked and the previous value of *debugger-hook*. When this happens, *debugger-hook* is bound to nil to prevent recursive errors. However, ANSI also mandates that *debugger-hook* not be invoked when the debugger is to be entered by the break function. For users who wish to provide an alternate debugger interface (and thus catch break entries into the debugger), SBCL provides sb-ext:*invoke-debugger-hook*, which is invoked during any entry into the debugger. - [variable] sb-ext:*invoke-debugger-hook* nil This is either nil or a designator for a function of two arguments, to be run when the debugger is about to be entered. The function is run with *invoke-debugger-hook* bound to nil to minimize recursive errors, and receives as arguments the condition that triggered debugger entry and the previous value of *invoke-debugger-hook*. This mechanism is an SBCL extension similar to the standard *debugger-hook*. In contrast to *debugger-hook*, it is observed by invoke-debugger even when called by break. ### Debugger Command Loop The debugger is an interactive read-eval-print loop much like the normal top level, but some symbols are interpreted as debugger commands instead of being evaluated. A debugger command starts with the symbol name of the command, possibly followed by some arguments on the same line. Some commands prompt for additional input. Debugger commands can be abbreviated by any unambiguous prefix: help can be typed as h, he, etc. The package is not significant in debugger commands; any symbol with the name of a debugger command will work. If you want to show the value of a variable that happens also to be the name of a debugger command you can wrap the variable in a progn to hide it from the command loop. The debugger prompt is ], where is the number of the current frame. Frames are numbered starting from zero at the top (most recent call), increasing down to the bottom. The current frame is the frame that commands refer to. It is possible to override the normal printing behaviour in the debugger by using the sb-ext:*debug-print-variable-alist*. - [variable] sb-ext:*debug-print-variable-alist* nil an association list describing new bindings for special variables to be used within the debugger. Eg. ((*print-length* . 10) (*print-level* . 6) (*print-pretty* . nil)) The variables in the car positions are bound to the values in the cdr during the execution of some debug commands. When evaluating arbitrary expressions in the debugger, the normal values of the printer control variables are in effect. Initially empty, *debug-print-variable-alist* is typically used to provide bindings for printer control variables. ### Stack Frames A stack frame is the run-time representation of a call to a function; the frame stores the state that a function needs to remember what it is doing. Frames have: - Variables (see Variable Access), which are the values being operated on. - Arguments to the call (which are really just particularly interesting variables). - A current source location (Source Location Printing), which is the place in the program where the function was running when it stopped to call another function, or because of an interrupt or error. #### Stack Motion These commands move to a new stack frame and print the name of the function and the values of its arguments in the style of a Lisp function call: - up: Move up to the next higher frame. More recent function calls are considered to be higher on the stack. - down: Move down to the next lower frame. - top: Move to the highest frame, that is, the frame where the debugger was entered. - bottom: Move to the lowest frame. - frame []: Move to the frame with the specified number. Prompts for the number if not supplied. The frame with number 0 is the frame where the debugger was entered. #### How Arguments are Printed A frame is printed to look like a function call, but with the actual argument values in the argument positions. So the frame for this call in the source: (myfun (+ 3 4) 'a) would look like this: (MYFUN 7 A) All keyword and optional arguments are displayed with their actual values; if the corresponding argument was not supplied, the value will be the default. So this call: (subseq "foo" 1) would look like this: (SUBSEQ "foo" 1 3) And this call: (string-upcase "test case") would look like this: (STRING-UPCASE "test case" :START 0 :END NIL) The arguments to a function call are displayed by accessing the argument variables. Although those variables are initialized to the actual argument values, they can be set inside the function; in this case the new value will be displayed. &rest arguments are handled somewhat differently. The value of the rest argument variable is displayed as the spread-out arguments to the call, so: (format t "~A is a ~A." "This" 'test) would look like this: (FORMAT T "~A is a ~A." "This" 'TEST) Rest arguments cause an exception to the normal display of keyword arguments in functions that have both &rest and &key arguments. In this case, the keyword argument variables are not displayed at all; the rest arg is displayed instead. So for these functions, only the keywords actually supplied will be shown, and the values displayed will be the argument values, not values of the (possibly modified) variables. If the variable for an argument is never referenced by the function, it will be deleted. The variable value is then unavailable, so the debugger prints # instead of the value. Similarly, if for any of a number of reasons the value of the variable is unavailable or not known to be available (Variable Access), then # will be printed instead of the argument value. Note that inline expansion and open-coding affect what frames are present in the debugger, see Debugger Policy Control. #### Function Names If a function is defined by defun it will appear in backtrace by that name. Functions defined by labels and flet will appear as (flet ) and (labels ) respectively. Anonymous lambdas will appear as (lambda ). ##### Entry Point Details Sometimes the compiler introduces new functions that are used to implement a user function, but are not directly specified in the source. This is mostly done for argument type and count checking. With recursive or block compiled functions, an additional external frame may appear before the frame representing the first call to the recursive function or entry to the compiled block. This is a consequence of the way the compiler works: there is nothing odd with your program. You may also see cleanup frames during the execution of unwind-protect cleanup code, and optional for variable argument entry points. #### Debug Tail Recursion The compiler is properly tail recursive. If a function call is in a tail recursive position, the stack frame will be deallocated at the time of the call, rather than after the call returns. Consider this backtrace: (BAR ...) (FOO ...) Because of tail recursion, it is not necessarily the case that foo directly called bar. It may be that foo called some other function foo2, which then called bar tail-recursively, as in this example: (defun foo () ... (foo2 ...) ...) (defun foo2 (...) ... (bar ...)) (defun bar (...) ...) Usually the elimination of tail-recursive frames makes debugging more pleasant, since these frames are mostly uninformative. If there is any doubt about how one function called another, it can usually be eliminated by finding the source location in the calling frame. See Source Location Printing. The elimination of tail-recursive frames can be prevented by disabling tail-recursion optimization, which happens when the debug optimization quality is greater than 2. See Debugger Policy Control. #### Unknown Locations and Interrupts The debugger operates using special debugging information attached to the compiled code. This debug information tells the debugger what it needs to know about the locations in the code where the debugger can be invoked. If the debugger somehow encounters a location not described in the debug information, then it is said to be unknown. If the code location for a frame is unknown, then some variables may be inaccessible, and the source location cannot be precisely displayed. There are three reasons why a code location could be unknown: - There is inadequate debug information due to the value of the debug optimization quality. See Debugger Policy Control. - The debugger was entered because of an interrupt such as C-c. - A hardware error such as a bus error occurred in code that was compiled unsafely due to the value of the safety optimization quality. In the last two cases, the values of argument variables are accessible, but may be incorrect. For more details on when variable values are accessible, see Variable Value Availability. It is possible for an interrupt to happen when a function call or return is in progress. The debugger may then flame out with some obscure error or insist that the bottom of the stack has been reached, when the real problem is that the current stack frame can't be located. If this happens, return from the interrupt and try again. ### Variable Access There are two ways to access the current frame's local variables in the debugger: list-locals and sb-debug:var. The debugger doesn't really understand lexical scoping; it has just one namespace for all the variables in the current stack frame. If a symbol is the name of multiple variables in the same function, then the reference appears ambiguous, even though lexical scoping specifies which value is visible at any given source location. If the scopes of the two variables are not nested, then the debugger can resolve the ambiguity by observing that only one variable is accessible. When there are ambiguous variables, the evaluator assigns each one a small integer identifier. The sb-debug:var function uses this identifier to distinguish between ambiguous variables. The list-locals command prints the identifier. In the following example, there are two variables named x. The first one has identifier 0 (which is not printed), the second one has identifier 1. X = 1 X#1 = 2 - list-locals []: This command prints the name and value of all variables in the current frame whose name has the specified , which may be a string or a symbol. If no is given, then all available variables are printed. If a variable has a potentially ambiguous name, then the name is printed with a # suffix, where is the small integer used to make the name unique. - [function] sb-debug:var name &optional (id 0) Return a variable's value if possible. name is a simple-string or symbol. If it is a simple-string, it is an initial substring of the variable's name. If name is a symbol, it has the same name and package as the variable whose value this function returns. If the symbol is uninterned, then the variable has the same name as the symbol, but it has no package. If name is the initial substring of variables with different names, then this returns no values after displaying the ambiguous names. If name determines multiple variables with the same name, then you must use the optional id argument to specify which one you want. If you left id unspecified, then this returns no values after displaying the distinguishing id values. The result of this function is limited to the availability of variable information. This is setfable. #### Variable Value Availability The value of a variable may be unavailable to the debugger in portions of the program where Lisp says that the variable is defined. If a variable value is not available, the debugger will not let you read or write that variable. With one exception, the debugger will never display an incorrect value for a variable. Rather than displaying incorrect values, the debugger tells you the value is unavailable. The one exception is this: if you interrupt (e.g. with C-c) or if there is an unexpected hardware error such as a bus error (which should only happen in unsafe code), then the values displayed for arguments to the interrupted frame might be incorrect. This exception applies only to the interrupted frame: any frame farther down the stack will be fine. > Note: Since the location of an interrupt or hardware error will > always be an unknown location, non-argument variable values will > never be available in the interrupted frame. See > Unknown Locations and Interrupts.) The value of a variable may be unavailable for these reasons: - The value of the debug optimization quality may have omitted debug information needed to determine whether the variable is available. Unless a variable is an argument, its value will only be available when debug is at least 2. - The compiler did lifetime analysis and determined that the value was no longer needed, even though its scope had not been exited. Lifetime analysis is inhibited when the debug optimization quality is 3. - The variable's name is an uninterned symbol (gensym). To save space, the compiler only dumps debug information about uninterned variables when the debug optimization quality is 3. - The frame's location is unknown (see Unknown Locations and Interrupts) because the debugger was entered due to an interrupt or unexpected hardware error. Under these conditions the values of arguments will be available, but might be incorrect. This is the exception mentioned above. - The variable (or the code referencing it) was optimized out of existence. Variables with no reads are always optimized away. The degree to which the compiler deletes variables will depend on the value of the compilation-speed optimization quality, but most source-level optimizations are done under all compilation policies. - The variable is never set and its definition looks like (LET ((var1 var2)) ...) In this case, var1 is substituted with var2. - The variable is never set and is referenced exactly once. In this case, the reference is substituted with the variable initial value. Since it is especially useful to be able to get the arguments to a function, argument variables are treated specially when the speed optimization quality is less than 3 and the debug quality is at least 1. With this compilation policy, the values of argument variables are almost always available everywhere in the function, even at unknown locations. For non-argument variables, debug must be at least 2 for values to be available, and even then, values are only available at known locations. #### Note On Lexical Variable Access When the debugger command loop establishes variable bindings for available variables, these variable bindings have lexical scope and dynamic extent. You can close over them, but such closures can't be used as upward function arguments. > Note: The variable bindings are actually created using the Lisp > symbol-macrolet special form. You can also set local variables using setq, but if the variable was closed over in the original source and never set, then setting the variable in the debugger may not change the value in all the functions the variable is defined in. Another risk of setting variables is that you may assign a value of a type that the compiler proved the variable could never take on. This may result in bad things happening. ### Source Location Printing One of the debugger's capabilities is source level debugging of compiled code. These commands display the source location for the current frame: - source []: This command displays the file that the current frame's function was defined from (if it was defined from a file), and then the source form responsible for generating the code that the current frame was executing. If is specified, then it is an integer specifying the number of enclosing levels of list structure to print. The source form for a location in the code is the innermost list present in the original source that encloses the form responsible for generating that code. If the actual source form is not a list, then some enclosing list will be printed. For example, if the source form was a reference to the variable *some-random-special*, then the innermost enclosing evaluated form will be printed. Here are some possible enclosing forms: (let ((a *some-random-special*)) ...) (+ *some-random-special* ...) If the code at a location was generated from the expansion of a macro or a source-level compiler optimization, then the form in the original source that expanded into that code will be printed. Suppose the file /usr/me/mystuff.lisp looked like this: (defmacro mymac () '(myfun)) (defun foo () (mymac) ...) If foo has called myfun, and is waiting for it to return, then the source command would print: ; File: /usr/me/mystuff.lisp (MYMAC) Note that the macro use was printed, not the actual function call form, (myfun). If enclosing source is printed by giving an argument to source or vsource, then the actual source form is marked by wrapping it in a list whose first element is #:***here***. In the previous example, source 1 would print: ; File: /usr/me/mystuff.lisp (DEFUN FOO () (#:***HERE*** (MYMAC)) ...) #### How the Source is Found If the code was defined from Lisp by compile or eval, then the source can always be reliably located. If the code was defined from a FASL file created by compile-file, then the debugger gets the source forms it prints by reading them from the original source file. This is a potential problem, since the source file might have moved or changed since the time it was compiled. The source file is opened using the truename of the source file pathname originally given to the compiler. This is an absolute pathname with all logical names and symbolic links expanded. If the file can't be located using this name, then the debugger gives up and signals an error. If the source file can be found, but has been modified since the time it was compiled, the debugger prints this warning: ; File has been modified since compilation: ; ; Using form offset instead of character position. where is the name of the source file. It then proceeds using a robust but not foolproof heuristic for locating the source. This heuristic works if: - No top-level forms before the top-level form containing the source have been added or deleted, and - the top-level form containing the source has not been modified much. (More precisely, none of the list forms beginning before the source form have been added or deleted.) If the heuristic doesn't work, the displayed source will be wrong, but will probably be near the actual source. If the "shape" of the top-level form in the source file is too different from the original form, then an error will be signaled. When the heuristic is used, the source location commands are noticeably slowed. Source location printing can also be confused if (after the source was compiled) a read-macro you used in the code was redefined to expand into something different, or if a read-macro ever returns the same eq list twice. If you don't define read macros and don't use ## in perverted ways, you don't need to worry about this. #### Source Location Availability Source location information is only available when the debug optimization quality is at least 2. If source location information is unavailable, the source commands will give an error message. If source location information is available, but the source location is unknown because of an interrupt or unexpected hardware error (see Unknown Locations and Interrupts), then the command will print Unknown location: using block start. and then proceed to print the source location for the start of the basic block enclosing the code location. It's a bit complicated to explain exactly what a basic block is, but here are some properties of the block start location: - The block start location may be the same as the true location. - The block start location will never be later in the program's flow of control than the true location. - No conditional control structures (such as if, cond, or) will intervene between the block start and the true location (but note that some conditionals present in the original source could be optimized away.) Function calls do not end basic blocks. - The head of a loop will be the start of a block. - The programming language concept of block structure and the Lisp block special form are totally unrelated to the compiler's basic block. In other words, the true location lies between the printed location and the next conditional (but watch out because the compiler may have changed the program on you.) ### Debugger Policy Control The compilation policy specified by optimize declarations affects the behavior seen in the debugger. The debug quality directly affects the debugger by controlling the amount of debugger information dumped. Other optimization qualities have indirect but observable effects due to changes in the way compilation is done. Unlike the other optimization qualities (which are compared in relative value to evaluate tradeoffs), the debug optimization quality is directly translated to a level of debug information. This absolute interpretation allows the user to count on a particular amount of debug information being available even when the values of the other qualities are changed during compilation. These are the levels of debug information that correspond to the values of the debug quality: - 0: Only the function name and enough information to allow the stack to be parsed. - > 0: Any level greater than 0 gives level 0 plus all argument variables. Values will only be accessible if the argument variable is never set and speed is not 3. SBCL allows any real value for optimization qualities. It may be useful to specify 0.5 to get backtrace argument display without argument documentation. - 1: Level 1 provides argument documentation (printed argument lists) and derived argument/result type information. This makes describe more informative, and allows the compiler to do compile-time argument count and type checking for any calls compiled at run-time. This is the default. - 2: Level 1 plus all interned local variables, source location information, and lifetime information that tells the debugger when arguments are available (even when speed is 3 or the argument is set). - > 2: Any level greater than 2 gives level 2 and in addition disables tail-call optimization, so that the backtrace will contain frames for all invoked functions, even those in tail positions. - 3: Level 2 plus all uninterned variables. In addition, lifetime analysis is disabled (even when speed is 3), ensuring that all variable values are available at any known location within the scope of the binding. This has a speed penalty in addition to the obvious space penalty. Inlining of local functions is inhibited so that they may be traced. - > (max speed space): If debug is greater than both speed and space, the command return can be used to continue execution by returning a value from the current stack frame. - > (max speed space compilation-speed): If debug is greater than all of speed, space and compilation-speed the code will be steppable (see Single Stepping). As you can see, if the speed quality is 3, debugger performance is degraded. This effect comes from the elimination of argument variable special-casing (see Variable Value Availability). Some degree of speed/debuggability tradeoff is unavoidable, but the effect is not too drastic when debug is at least 2. In addition to inline and notinline declarations, the relative values of the speed and space qualities also change whether functions are inline expanded. If a function is inline expanded, then there will be no frame to represent the call, and the arguments will be treated like any other local variable. Functions may also be semi inline, in which case there is a frame to represent the call, but the call is to an optimized local version of the function, not to the original function. ### Exiting Commands These commands get you out of the debugger. - toplevel: Throw to top level. - restart []: Invoke the th restart case as displayed by the error command. If is not specified, the available restart cases are reported. - continue: Call continue on the condition given to debug. If there is no restart case named continue, then an error is signaled. - abort: Call abort on the condition given to debug. This is useful for popping debug command loop levels or aborting to top level, as the case may be. - return : Return value from the current stack frame. This command is available when the debug optimization quality is greater than both speed and space. Care must be taken that the value is of the same type as SBCL expects the stack frame to return. - restart-frame: Restart execution of the current stack frame. This command is available when the debug optimization quality is greater than both speed and space and when the frame is for a global function. If the function is redefined in the debugger before the frame is restarted, the new function will be used. ### Information Commands Most of these commands print information about the current frame or function, but a few show general information. - help or ?: Display a synopsis of debugger commands. - describe: Call describe on the current function and displays the number of local variables. - print: Display the current function call as it would be displayed by moving to this frame. - error: Print the condition given to invoke-debugger and the active proceed cases. - backtrace []: Display all the frames from the current to the bottom. Only shows frames if specified. The printing is controlled by sb-debug:*debug-print-variable-alist*. ### Breakpoint Commands SBCL supports setting of breakpoints inside compiled functions and stepping of compiled code. Breakpoints can only be set at known locations (see Unknown Locations and Interrupts), so these commands are largely useless unless the debug optimize quality is at least 2 (see Debugger Policy Control). These commands manipulate breakpoints: - breakpoint [