Next: Introduction [Contents][Index]
SLIME is the “Superior Lisp Interaction Mode for Emacs”. This is the manual for version 2.30. (Last updated November 21, 2024)
Written by Luke Gorrie and others.
This file has been placed in the public domain.
inferior-slime-modeslime-autodoc-modeslime-sprofslime-fancyNext: Getting started, Previous: SLIME, Up: SLIME [Contents][Index]
SLIME is the “Superior Lisp Interaction Mode for Emacs.”
SLIME extends Emacs with support for interactive programming in
Common Lisp. The features are centered around slime-mode, an
Emacs minor-mode that complements the standard lisp-mode. While
lisp-mode supports editing Lisp source files, slime-mode
adds support for interacting with a running Common Lisp process for
compilation, debugging, documentation lookup, and so on.
The slime-mode programming environment follows the example of
Emacs’s native Emacs Lisp environment. We have also included good
ideas from similar systems (such as ILISP) and some new
ideas of our own.
SLIME is constructed from two parts: a user-interface written in Emacs Lisp, and a supporting server program written in Common Lisp. The two sides are connected together with a socket and communicate using an RPC-like protocol.
The Lisp server is primarily written in portable Common Lisp. The required implementation-specific functionality is specified by a well-defined interface and implemented separately for each Lisp implementation. This makes SLIME readily portable.
Next: Using Slime mode, Previous: Introduction, Up: SLIME [Contents][Index]
This chapter tells you how to get SLIME up and running.
Next: Downloading SLIME, Up: Getting started [Contents][Index]
SLIME supports a wide range of operating systems and Lisp implementations. SLIME runs on Unix systems, Mac OSX, and Microsoft Windows. GNU Emacs versions 24.3 and above are supported. XEmacs is not supported anymore.
The supported Lisp implementations, roughly ordered from the best-supported, are:
Most features work uniformly across implementations, but some are prone to variation. These include the precision of placing compiler-note annotations, XREF support, and fancy debugger commands (like “restart frame”).
Next: Installation, Previous: Supported Platforms, Up: Getting started [Contents][Index]
You can choose between using a released version of SLIME or accessing our Git repository directly. You can download the latest released version from our website:
http://github.com/slime/slime/
We recommend that users who participate in the slime-devel
mailing list use the Git version of the code.
Next: Git incantations, Up: Downloading SLIME [Contents][Index]
SLIME is available from the Git repository on
github.com. You have the option to use either the very latest
code or the tagged FAIRLY-STABLE snapshot.
The latest version tends to have more features and fewer bugs than the
FAIRLY-STABLE version, but it can be unstable during times of
major surgery. As a rule-of-thumb recommendation we suggest that if
you follow the slime-devel mailing list then you’re better off
with the latest version (we’ll send a note when it’s undergoing major
hacking). If you don’t follow the mailing list you won’t know the
status of the latest code, so tracking FAIRLY-STABLE or using a
released version is the safe option.
If you download from Git then remember to git pull
occasionally. Improvements are continually being committed, and the
FAIRLY-STABLE tag is moved forward from time to time.
Previous: Downloading from Git, Up: Downloading SLIME [Contents][Index]
To download the very latest SLIME you first configure
your GitROOT and login to the repository.
git clone https://github.com/slime/slime.git
You might substitute https for http if you’re having
problems with that protocol.
If you want to hack on SLIME, use Github’s fork functionality and submit a pull request. Be sure to first read the CONTRIBUTING.md file first.
Next: Running SLIME, Previous: Downloading SLIME, Up: Getting started [Contents][Index]
The easiest way to install and keep SLIME up-to-date is using Emacs’s built-in package manager. SLIME is available from the MELPA repository. After setting up the MELPA repository, SLIME can be installed via M-x package-install RET slime RET. You should then define your default Lisp in your .emacs as follows:
(setq inferior-lisp-program "/opt/sbcl/bin/sbcl")
At this point, you should be ready to start running SLIME.
This is the minimal configuration with the fewest frills. If the basic setup is working, you can try additional modules (Loading Contrib Packages).
If you’d rather install SLIME directly from its git repository, you will need to add a few extra lines in your .emacs:
;; Setup load-path, autoloads and your lisp system ;; Not needed if you install SLIME via MELPA (add-to-list 'load-path "~/dir/to/cloned/slime") (require 'slime-autoloads) (setq inferior-lisp-program "/opt/sbcl/bin/sbcl")
You may optionally byte-compile SLIME using make compile
contrib-compile.
Next: Setup Tuning, Previous: Installation, Up: Getting started [Contents][Index]
SLIME is started with the Emacs command M-x slime. This uses
the inferior-lisp package to start a Lisp process, loads and
starts the Lisp-side server (known as “Swank”), and establishes a
socket connection between Emacs and Lisp. Finally a REPL buffer is
created where you can enter Lisp expressions for evaluation.
At this point SLIME is up and running and you can start exploring.
Previous: Running SLIME, Up: Getting started [Contents][Index]
This section explains ways to perform basic extensions to SLIME, and how to configure SLIME for multiple Lisp systems and how to reduce SLIME’s startup time.
Please proceed with this section only if your basic setup works. If you are happy with the basic setup, skip this section.
For contrib modules see Loading Contrib Packages.
Next: Multiple Lisps, Up: Setup Tuning [Contents][Index]
Once you have the basic no-frills setup working, you can enhance your SLIME installation with bundled extensions:
;; Setup load-path, autoloads and your lisp system (add-to-list 'load-path "~/dir/to/cloned/slime") (require 'slime-autoloads)
See see Loading Contrib Packages for more information on SLIME’s contrib system.
To customize a particular binding in one of SLIME’s keymaps, you can add one of the following to your init file:
(add-hook 'slime-load-hook
(lambda ()
(define-key slime-prefix-map (kbd "M-h") 'slime-documentation-lookup)))
The former technique works only for SLIME’s core keymaps, not it’s contribs’. For those you can use the latter form which works for any Emacs library. See also see Customization for more advanced configuration options.
Next: Loading Swank faster, Previous: Basic customization, Up: Setup Tuning [Contents][Index]
By default, the command M-x slime starts the program specified
with inferior-lisp-program. If you invoke M-x slime with
a prefix argument, Emacs prompts for the program which should be
started instead. If you need that frequently or if the command
involves long filenames it’s more convenient to set the
slime-lisp-implementations variable in your .emacs. For
example here we define two programs:
(setq slime-lisp-implementations
'((cmucl ("cmucl" "-quiet"))
(sbcl ("/opt/sbcl/bin/sbcl") :coding-system utf-8-unix)))
This variable holds a list of programs and if you invoke SLIME with
a negative prefix argument, M-- M-x slime, you can select a
program from that list. When called without a prefix, either the name
specified in slime-default-lisp, or the first item of the list will be used.
The elements of the list should look like
(NAME (PROGRAM PROGRAM-ARGS...) &key CODING-SYSTEM INIT INIT-FUNCTION ENV)
NAMEis a symbol and is used to identify the program.
PROGRAMis the filename of the program. Note that the filename can contain spaces.
PROGRAM-ARGSis a list of command line arguments.
CODING-SYSTEMthe coding system for the connection. (see slime-net-coding-system)x
INITshould be a function which takes two arguments: a filename and a
character encoding. The function should return a Lisp expression as a
string which instructs Lisp to start the Swank server and to write the
port number to the file. At startup, SLIME starts the Lisp process
and sends the result of this function to Lisp’s standard input. As
default, slime-init-command is used. An example is shown in
Loading Swank faster.
INIT-FUNCTIONshould be a function which takes no arguments. It is called after the connection is established. (See also slime-connected-hook.)
ENVspecifies a list of environment variables for the subprocess. E.g.
(sbcl-cvs ("/home/me/sbcl-cvs/src/runtime/sbcl"
"--core" "/home/me/sbcl-cvs/output/sbcl.core")
:env ("SBCL_HOME=/home/me/sbcl-cvs/contrib/"))
initializes SBCL_HOME in the subprocess.
Previous: Multiple Lisps, Up: Setup Tuning [Contents][Index]
For SBCL, we recommend that you create a custom core file with socket support and POSIX bindings included because those modules take the most time to load. To create such a core, execute the following steps:
shell$ sbcl * (mapc 'require '(sb-bsd-sockets sb-posix sb-introspect sb-cltl2 asdf)) * (save-lisp-and-die "sbcl.core-for-slime")
After that, add something like this to your .emacs:
(setq slime-lisp-implementations
'((sbcl ("sbcl" "--core" "sbcl.core-for-slime"))))
For maximum startup speed you can include the Swank server directly in a core file. The disadvantage of this approach is that the setup is a bit more involved and that you need to create a new core file when you want to update SLIME or SBCL. The steps to execute are:
shell$ sbcl * (load ".../slime/swank-loader.lisp") * (swank-loader:dump-image "sbcl.core-with-swank")
Then add this to your .emacs:
(setq slime-lisp-implementations
'((sbcl ("sbcl" "--core" "sbcl.core-with-swank")
:init (lambda (port-file _)
(format "(swank:start-server %S)\n" port-file)))))
Similar setups should also work for other Lisp implementations.
Next: SLDB: the SLIME debugger, Previous: Getting started, Up: SLIME [Contents][Index]
SLIME’s commands are provided via slime-mode, a minor-mode
used in conjunction with Emacs’s lisp-mode. This chapter
describes the slime-mode and its relatives.
Next: Evaluation commands, Up: Using Slime mode [Contents][Index]
To use SLIME comfortably it is important to understand a few “global” user-interface characteristics. The most important principles are described in this section.
Next: *inferior-lisp* buffer, Up: User-interface conventions [Contents][Index]
Some SLIME commands create temporary buffers to display their results. Although these buffers usually have their own special-purpose major-modes, certain conventions are observed throughout.
Temporary buffers can be dismissed by pressing q. This kills the
buffer and restores the window configuration as it was before the
buffer was displayed. Temporary buffers can also be killed with the
usual commands like kill-buffer, in which case the previous
window configuration won’t be restored.
Pressing RET is supposed to “do the most obvious useful thing.” For instance, in an apropos buffer this prints a full description of the symbol at point, and in an XREF buffer it displays the source code for the reference at point. This convention is inherited from Emacs’s own buffers for apropos listings, compilation results, etc.
Temporary buffers containing Lisp symbols use slime-mode in
addition to any special mode of their own. This makes the usual
SLIME commands available for describing symbols, looking up
function definitions, and so on.
Initial focus of those “description” buffers depends on the variable
slime-description-autofocus. If nil (the default),
description buffers do not receive focus automatically, and vice
versa.
Next: Multithreading, Previous: Temporary buffers, Up: User-interface conventions [Contents][Index]
*inferior-lisp* bufferSLIME internally uses the comint package to start Lisp
processes. This has a few user-visible consequences, some good and
some not-so-terribly. To avoid confusion it is useful to understand
the interactions.
The buffer *inferior-lisp* contains the Lisp process’s own
top-level. This direct access to Lisp is useful for troubleshooting,
and some degree of SLIME integration is available using the
inferior-slime-mode. Many people load the better integrated SLIME
REPL contrib module (see REPL: the “top level”) and ignore
the *inferior-lisp* buffer. (see Loading Contrib Packages for
information on how to enable the REPL.)
Next: Key bindings, Previous: *inferior-lisp* buffer, Up: User-interface conventions [Contents][Index]
If the Lisp system supports multithreading, SLIME spawns a new thread for each request, e.g., C-x C-e creates a new thread to evaluate the expression. An exception to this rule are requests from the REPL: all commands entered in the REPL buffer are evaluated in a dedicated REPL thread.
Some complications arise with multithreading and special variables.
Non-global special bindings are thread-local, e.g., changing the value
of a let bound special variable in one thread has no effect on the
binding of the variables with the same name in other threads. This
makes it sometimes difficult to change the printer or reader behaviour
for new threads. The variable
swank:*default-worker-thread-bindings* was introduced for such
situations: instead of modifying the global value of a variable, add a
binding the swank:*default-worker-thread-bindings*. E.g., with
the following code, new threads will read floating point values as
doubles by default:
(push '(*read-default-float-format* . double-float)
swank:*default-worker-thread-bindings*).
Previous: Multithreading, Up: User-interface conventions [Contents][Index]
In general we try to make our key bindings fit with the overall Emacs
style. We also have the following somewhat unusual convention of our
own: when entering a three-key sequence, the final key can be pressed
either with control or unmodified. For example, the
slime-describe-symbol command is bound to C-c C-d d, but
it also works to type C-c C-d C-d. We’re simply binding both key
sequences because some people like to hold control for all three keys
and others don’t, and with the two-key prefix we’re not afraid of
running out of keys.
There is one exception to this rule, just to trip you up. We never bind C-h anywhere in a key sequence, so C-c C-d C-h doesn’t do the same thing as C-c C-d h. This is because Emacs has a built-in default so that typing a prefix followed by C-h will display all bindings starting with that prefix, so C-c C-d C-h will actually list the bindings for all documentation commands. This feature is just a bit too useful to clobber!
“Are you deliberately spiting Emacs’s brilliant online help facilities? The gods will be angry!”
This is a brilliant piece of advice. The Emacs online help facilities are your most immediate, up-to-date and complete resource for keybinding information. They are your friends:
describe-key
“What does this key do?”
Describes current function bound to <key> for focus buffer.
describe-bindings
“Exactly what bindings are available?”
Lists the current key-bindings for the focus buffer.
describe-mode
“Tell me all about this mode”
Shows all the available major mode keys, then the minor mode keys, for
the modes of the focus buffer.
view-lossage
“Woah, what key chord did I just do?”
Shows you the literal sequence of keys you’ve pressed in order.
Note: In this documentation the designation C-h is a
canonical key which might actually mean Ctrl-h, or F1, or
whatever you have help-command bound to in your
.emacs. Here is a common situation:
(global-set-key [f1] 'help-command) (global-set-key "\C-h" 'delete-backward-char)
In this situation everywhere you see C-h in the documentation you would substitute F1.
You can assign or change default key bindings globally using the
global-set-key function in your ~/.emacs file like this:
(global-set-key "\C-c s" 'slime-selector)
which binds C-c s to the function slime-selector.
Alternatively, if you want to assign or change a key binding in just a
particular slime mode, you can use the define-key function
in your ~/.emacs file like this:
(define-key slime-repl-mode-map (kbd "C-c ;")
'slime-insert-balanced-comments)
which binds C-c ; to the function
slime-insert-balanced-comments in the REPL buffer.
Next: Compilation commands, Previous: User-interface conventions, Up: Using Slime mode [Contents][Index]
These commands each evaluate a Common Lisp expression in a different way. Usually they mimic commands for evaluating Emacs Lisp code. By default they show their results in the echo area, but a prefix argument causes the results to be inserted in the current buffer.
Evaluate the expression before point and show the result in the echo area.
Evaluate the current toplevel form and show the result in the echo area. ‘C-M-x’ treats ‘defvar’ expressions specially. Normally, evaluating a ‘defvar’ expression does nothing if the variable it defines already has a value. But ‘C-M-x’ unconditionally resets the variable to the initial value specified in the ‘defvar’ expression. This special feature is convenient for debugging Lisp programs.
If C-M-x or C-x C-e is given a numeric argument, it inserts the value into the current buffer, rather than displaying it in the echo area.
Evaluate an expression read from the minibuffer.
Evaluate the region.
Evaluate the expression before point and pretty-print the result in a fresh buffer.
Edit the value of a setf-able form in a new buffer *Edit <form>*. The value is inserted into a temporary buffer for editing and then set in Lisp when committed with C-c C-c.
Undefine the function, with fmakunbound, for the symbol at
point.
Next: Completion commands, Previous: Evaluation commands, Up: Using Slime mode [Contents][Index]
SLIME has fancy commands for compiling functions, files, and packages. The fancy part is that notes and warnings offered by the Lisp compiler are intercepted and annotated directly onto the corresponding expressions in the Lisp source buffer. (Give it a try to see what this means.)
Compile the top-level form at point. The region blinks shortly to give some feedback which part was chosen.
With (positive) prefix argument the form is compiled with maximal debug settings (C-u C-c C-c). With negative prefix argument it is compiled for speed (M-- C-c C-c). If a numeric argument is passed set debug or speed settings to it depending on its sign.
The code for the region is executed after compilation. In principle, the command writes the region to a file, compiles that file, and loads the resulting code.
Compile and load the current buffer’s source file. If the compilation step fails, the file is not loaded. It’s not always easy to tell whether the compilation failed: occasionally you may end up in the debugger during the load step.
With (positive) prefix argument the file is compiled with maximal debug settings (C-u C-c C-k). With negative prefix argument it is compiled for speed (M-- C-c C-k). If a numeric argument is passed set debug or speed settings to it depending on its sign.
Compile (but don’t load) the current buffer’s source file.
Load a Lisp file. This command uses the Common Lisp LOAD function.
Compile the selected region.
The annotations are indicated as underlining on source forms. The compiler message associated with an annotation can be read either by placing the mouse over the text or with the selection commands below.
Move the point to the next compiler note and displays the note.
Move the point to the previous compiler note and displays the note.
Remove all annotations from the buffer.
Visit the next-error message. This is not actually a SLIME command but SLIME creates a hidden buffer so that most of the Compilation mode commands (See (emacs)Compilation Mode) work similarly for Lisp as for batch compilers.
Next: Finding definitions (“Meta-Point” commands)., Previous: Compilation commands, Up: Using Slime mode [Contents][Index]
Completion commands are used to complete a symbol or form based on what is already present at point. Classical completion assumes an exact prefix and gives choices only where branches may occur. Fuzzy completion tries harder.
Complete the symbol at point. Note that three styles of completion are available in SLIME; the default is similar to normal Emacs completion (see slime-completion-at-point-functions).
Next: Documentation commands, Previous: Completion commands, Up: Using Slime mode [Contents][Index]
The familiar M-. command is provided. For generic functions this
command finds all methods, and with some systems it does other fancy
things (like tracing structure accessors to their DEFSTRUCT
definition).
Go to the definition of the symbol at point.
Go back to the point where M-. was invoked. This gives multi-level backtracking when M-. has been used several times.
Like slime-edit-definition but switches to the other window to
edit the definition in.
Like slime-edit-definition but opens another frame to edit the
definition in.
Use an ETAGS table to find definition at point.
Next: Cross-reference commands, Previous: Finding definitions (“Meta-Point” commands)., Up: Using Slime mode [Contents][Index]
SLIME’s online documentation commands follow the example of Emacs Lisp. The commands all share the common prefix C-c C-d and allow the final key to be modified or unmodified (see Key bindings.)
The space key inserts a space, but also looks up and displays the argument list for the function at point, if there is one.
Describe the symbol at point.
Describe the function at point.
Perform an apropos search on Lisp symbol names for a regular expression match and display their documentation strings. By default the external symbols of all packages are searched. With a prefix argument you can choose a specific package and whether to include unexported symbols.
Like slime-apropos but also includes internal symbols by default.
Show apropos results of all symbols in a package. This command is for browsing a package at a high-level. With package-name completion it also serves as a rudimentary Smalltalk-ish image-browser.
Lookup the symbol at point in the Common Lisp Hyperspec. This
uses the familiar hyperspec.el to show the appropriate section
in a web browser. The Hyperspec is found either on the Web or in
common-lisp-hyperspec-root, and the browser is selected by
browse-url-browser-function.
Note: this is one case where C-c C-d h is not the same as C-c C-d C-h.
Lookup a format character in the Common Lisp Hyperspec.
Lookup a reader macro in the Common Lisp Hyperspec.
Next: Macro-expansion commands, Previous: Documentation commands, Up: Using Slime mode [Contents][Index]
SLIME’s cross-reference commands are based on the support provided by the Lisp system, which varies widely between Lisps. For systems with no built-in XREF support SLIME queries a portable XREF package, which is taken from the CMU AI Repository and bundled with SLIME.
Each command operates on the symbol at point, or prompts if there is none. With a prefix argument they always prompt. You can either enter the key bindings as shown here or with the control modified on the last key, See Key bindings.
Show function callers.
Show all known callees.
Show references to global variable.
Show bindings of a global variable.
Show assignments to a global variable.
Show expansions of a macro.
Show all known methods specialized on a class.
There are also “List callers/callees” commands. These operate by rummaging through function objects on the heap at a low-level to discover the call graph. They are only available with some Lisp systems, and are most useful as a fallback when precise XREF information is unavailable.
List callers of a function.
List callees of a function.
Commands available in Xref buffers
Show definition at point in the other window. Do not leave Xref buffer.
Show definition at point in the other window and close Xref buffer.
Recompile definition at point.
Recompile all definitions.
Next: Disassembly commands, Previous: Cross-reference commands, Up: Using Slime mode [Contents][Index]
Macroexpand (or compiler-macroexpand) the expression starting at point once. If invoked with a prefix argument, use macroexpand instead of macroexpand-1 (or compiler-macroexpand instead of compiler-macroexpand-1).
Macroexpand the expression starting at point once. If invoked with a prefix argument, use macroexpand instead of macroexpand-1.
Fully macroexpand the expression starting at point.
Display the compiler-macro expansion of sexp starting at point.
Repeatedly expand compiler macros of sexp starting at point.
For additional minor-mode commands and discussion, see slime-macroexpansion-minor-mode.
Next: Abort/Recovery commands, Previous: Macro-expansion commands, Up: Using Slime mode [Contents][Index]
Disassemble the function definition of the symbol at point.
Toggle tracing of the function at point. If invoked with a prefix argument, read additional information, like which particular method should be traced.
Untrace all functions.
Next: Inspector commands, Previous: Disassembly commands, Up: Using Slime mode [Contents][Index]
Interrupt Lisp (send SIGINT).
Restart the inferior-lisp process.
Synchronize the current package and working directory from Emacs to Lisp.
Set the current package of the REPL.
Set the current directory of the Lisp process. This also changes the current directory of the REPL buffer.
Print the current directory of the Lisp process.
Next: Profiling commands, Previous: Abort/Recovery commands, Up: Using Slime mode [Contents][Index]
The SLIME inspector is a Emacs-based alternative to the
standard INSPECT function. The inspector presents objects in
Emacs buffers using a combination of plain text, hyperlinks to related
objects.
The inspector can easily be specialized for the objects in your own
programs. For details see the inspect-for-emacs generic
function in swank/backend.lisp.
Inspect the value of an expression entered in the minibuffer.
The standard commands available in the inspector are:
If point is on a value then recursively call the inspector on that value. If point is on an action then call that action.
Describe the slot at point.
Evaluate an expression in the context of the inspected object. The
variable * will be bound to the inspected object.
Toggle between verbose and terse mode. Default is determined by ‘swank:*inspector-verbose*’.
Go back to the previous object (return from RET).
The inverse of l. Also bound to SPC.
Reinspect.
Dismiss the inspector buffer.
Pretty print in another buffer object at point.
Find source of object at point.
Fetch all inspector contents and go to the end.
Store the value under point in the variable ‘*’. This can then be used to access the object in the REPL.
Jump to the next and previous inspectable object respectively.
Next: Shadowed Commands, Previous: Inspector commands, Up: Using Slime mode [Contents][Index]
The profiling commands are based on CMUCL’s profiler. These are simple wrappers around functions which usually print something to the output buffer.
Toggle profiling of a function.
Profile all functions in a package.
Profile all functions which names contain a substring.
Unprofile all functions.
Report profiler data.
Reset profiler data.
Show list of currently profiled functions.
Next: Semantic indentation, Previous: Profiling commands, Up: Using Slime mode [Contents][Index]
This key-binding is shadowed from inf-lisp.
Next: Reader conditional fontification, Previous: Shadowed Commands, Up: Using Slime mode [Contents][Index]
SLIME automatically discovers how to indent the macros in your Lisp
system. To do this the Lisp side scans all the macros in the system and
reports to Emacs all the ones with &body arguments. Emacs then
indents these specially, putting the first arguments four spaces in and
the “body” arguments just two spaces, as usual.
This should “just work.” If you are a lucky sort of person you needn’t read the rest of this section.
To simplify the implementation, SLIME doesn’t distinguish between macros with the same symbol-name but different packages. This makes it fit nicely with Emacs’s indentation code. However, if you do have several macros with the same symbol-name then they will all be indented the same way, arbitrarily using the style from one of their arglists. You can find out which symbols are involved in collisions with:
(swank:print-indentation-lossage)
If a collision causes you irritation, don’t have a nervous breakdown,
just override the Elisp symbol’s common-lisp-indent-function
property to your taste. SLIME won’t override your custom settings, it
just tries to give you good defaults.
A more subtle issue is that imperfect caching is used for the sake of performance. 1
In an ideal world, Lisp would automatically scan every symbol for indentation changes after each command from Emacs. However, this is too expensive to do every time. Instead Lisp usually just scans the symbols whose home package matches the one used by the Emacs buffer where the request comes from. That is sufficient to pick up the indentation of most interactively-defined macros. To catch the rest we make a full scan of every symbol each time a new Lisp package is created between commands – that takes care of things like new systems being loaded.
You can use M-x slime-update-indentation to force all symbols to be scanned for indentation information.
Previous: Semantic indentation, Up: Using Slime mode [Contents][Index]
SLIME automatically evaluates reader-conditional expressions, like
#+linux, in source buffers and “grays out” code that will be
skipped for the current Lisp connection.
Next: Misc, Previous: Using Slime mode, Up: SLIME [Contents][Index]
SLIME has a custom Emacs-based debugger called SLDB. Conditions
signalled in the Lisp system invoke SLDB in Emacs by way of the
Lisp *DEBUGGER-HOOK*.
SLDB pops up a buffer when a condition is signalled. The buffer displays a description of the condition, a list of restarts, and a backtrace. Commands are offered for invoking restarts, examining the backtrace, and poking around in stack frames.
Next: Invoking restarts, Up: SLDB: the SLIME debugger [Contents][Index]
Commands for examining the stack frame at point.
Toggle display of local variables and CATCH tags.
View the frame’s current source expression. The expression is presented in the Lisp source file’s buffer.
Evaluate an expression in the frame. The expression can refer to the available local variables in the frame.
Evaluate an expression in the frame and pretty-print the result in a temporary buffer.
Disassemble the frame’s function. Includes information such as the instruction pointer within the frame.
Inspect the result of evaluating an expression in the frame.
Recompile frame. C-u C-c C-c for recompiling with maximum debug settings.
Next: Navigating between frames, Previous: Examining frames, Up: SLDB: the SLIME debugger [Contents][Index]
Invoke the ABORT restart.
“Quit” – For SLIME evaluation requests, invoke a restart which restores to a known program state. For errors in other threads, see *SLDB-QUIT-RESTART*.
Invoke the CONTINUE restart.
Invoke a restart by number.
Restarts can also be invoked by pressing RET or Mouse-2 on them in the buffer.
Next: Miscellaneous Commands, Previous: Navigating between frames, Up: SLDB: the SLIME debugger [Contents][Index]
Stepping is not available in all implementations and works very differently in those in which it is available.
Step to the next expression in the frame. For CMUCL that means, set a breakpoint at all those code locations in the current code block which are reachable from the current code location.
Step to the next form in the current function.
Stop single-stepping temporarily, but resume it once the current function returns.
Previous: Stepping, Up: SLDB: the SLIME debugger [Contents][Index]
Restart execution of the frame with the same arguments it was originally called with. (This command is not available in all implementations.)
Return from the frame with a value entered in the minibuffer. (This command is not available in all implementations.)
Exit SLDB and debug the condition using the Lisp system’s default debugger.
Inspect the condition currently being debugged.
Evaluate an expression entered in the minibuffer.
Attach debugger (e.g. gdb) to the current lisp process.
Next: Customization, Previous: SLDB: the SLIME debugger, Up: SLIME [Contents][Index]
Next: slime-macroexpansion-minor-mode, Up: Misc [Contents][Index]
slime-selectorThe slime-selector command is for quickly switching to
important buffers: the REPL, SLDB, the Lisp source you were just
hacking, etc. Once invoked the command prompts for a single letter to
specify which buffer it should display. If a prefix argument is
specified then the buffer is shown in the other window. Here are some
of the main options:
A help buffer listing all slime-selectors’s available buffers.
This includes some options of lesser importance, e.g. intended for
debugging SLIME, which are not included in this list.
The REPL buffer for the current SLIME connection.
The most recently activated SLDB buffer for the current connection.
The most recently visited lisp-mode source buffer.
The *slime-scratch* buffer (see slime-scratch).
SLIME connections buffer (see Multiple connections).
Cycle to the next Lisp connection (see Multiple connections).
SLIME threads buffer (see Multiple connections).
The *inferior-lisp* buffer for the current connection.
Show buffer in other window. Equivalent to supplying a prefix argument. Prompts again for which buffer to select.
slime-selector doesn’t have a key binding by default but we
suggest that you assign it a global one. You can bind it to C-c s
like this:
(global-set-key "\C-cs" 'slime-selector)
And then you can switch to the REPL from anywhere with C-c s r.
The macro def-slime-selector-method can be used to define new
buffers for slime-selector to find.
Next: Multiple connections, Previous: slime-selector, Up: Misc [Contents][Index]
Within a slime macroexpansion buffer some extra commands are provided (these commands are always available but are only bound to keys in a macroexpansion buffer).
Just like slime-macroexpand-1 but the original form is replaced with the expansion.
The last macroexpansion is performed again, the current contents of the macroexpansion buffer are replaced with the new expansion.
Close the expansion buffer.
Undo last macroexpansion operation.
Previous: slime-macroexpansion-minor-mode, Up: Misc [Contents][Index]
SLIME is able to connect to multiple Lisp processes at the same time. The M-x slime command, when invoked with a prefix argument, will offer to create an additional Lisp process if one is already running. This is often convenient, but it requires some understanding to make sure that your SLIME commands execute in the Lisp that you expect them to.
Some buffers are tied to specific Lisp processes. Each Lisp connection has its own REPL buffer, and all expressions entered or SLIME commands invoked in that buffer are sent to the associated connection. Other buffers created by SLIME are similarly tied to the connections they originate from, including SLDB buffers, apropos result listings, and so on. These buffers are the result of some interaction with a Lisp process, so commands in them always go back to that same process.
Commands executed in other places, such as slime-mode source
buffers, always use the “default” connection. Usually this is the
most recently established connection, but this can be reassigned via
the “connection list” buffer:
Pop up a buffer listing the established connections. It is also
available by the typing c from the SLIME selector
(slime-selector).
Change current Lisp connection by cycling through all connections. It
is also available by the typing n from the SLIME selector
(slime-selector).
Pop up a buffer listing the current threads. It is also available by
the typing t from the SLIME selector (slime-selector).
The buffer displayed by slime-list-connections gives a one-line
summary of each connection. The summary shows the connection’s serial
number, the name of the Lisp implementation, and other details of the
Lisp process. The current “default” connection is indicated with an
asterisk.
The commands available in the connection-list buffer are:
Pop to the REPL buffer of the connection at point.
Make the connection at point the “default” connection. It will then
be used for commands in slime-mode source buffers.
Update the connection list in the buffer.
Quit the connection list (kill buffer, restore window configuration).
Restart the Lisp process for the connection at point.
Connect to a running Swank server.
Disconnect all connections.
Abort the current attempt to connect.
Next: Tips and Tricks, Previous: Misc, Up: SLIME [Contents][Index]
Next: Lisp-side (Swank), Up: Customization [Contents][Index]
The Emacs part of SLIME can be configured with the Emacs
customize system, just use M-x customize-group slime
RET. Because the customize system is self-describing, we only cover a
few important or obscure configuration options here in the manual.
slime-truncate-linesThe value to use for truncate-lines in line-by-line summary
buffers popped up by SLIME. This is t by default, which
ensures that lines do not wrap in backtraces, apropos listings, and so
on. It can however cause information to spill off the screen.
slime-completion-at-point-functionsA list of functions used for completion of Lisp symbols. This works
as the standard
completion-at-point-functions
(see (elisp)Completion in Buffers). Three completion
styles are available: slime-simple-completion-at-point,
slime-complete-symbol* (see Compound Completion),
and slime-fuzzy-complete-symbol (see Fuzzy Completion).
The default is slime-simple-completion-at-point, which
completes in the usual Emacs way.
slime-filename-translationsThis variable controls filename translation between Emacs and the Lisp system. It is useful if you run Emacs and Lisp on separate machines which don’t share a common file system or if they share the filesystem but have different layouts, as is the case with SMB-based file sharing.
slime-net-coding-systemIf you want to transmit Unicode characters between Emacs and the Lisp system, you should customize this variable. E.g., if you use SBCL, you can set:
(setq slime-net-coding-system 'utf-8-unix)
To actually display Unicode characters you also need appropriate
fonts, otherwise the characters will be rendered as hollow boxes. If
you are using Allegro CL and GNU Emacs, you can also
use emacs-mule-unix as coding system. GNU Emacs has often
nicer fonts for the latter encoding. (Different encodings can be used
for different Lisps, see Multiple Lisps.)
Up: Emacs-side [Contents][Index]
slime-mode-hookThis hook is run each time a buffer enters slime-mode. It is
most useful for setting buffer-local configuration in your Lisp source
buffers. An example use is to enable slime-autodoc-mode
(see slime-autodoc-mode).
slime-connected-hookThis hook is run when SLIME establishes a connection to a Lisp server. An example use is to create a Typeout frame (See Typeout frames.)
sldb-hookThis hook is run after SLDB is invoked. The hook functions are
called from the SLDB buffer after it is initialized. An example use
is to add sldb-print-condition to this hook, which makes all
conditions debugged with SLDB be recorded in the REPL buffer.
Previous: Emacs-side, Up: Customization [Contents][Index]
The Lisp server side of SLIME (known as “Swank”) offers several variables to configure. The initialization file ~/.swank.lisp is automatically evaluated at startup and can be used to set these variables.
Next: Other configurables, Up: Lisp-side (Swank) [Contents][Index]
The most important configurable is SWANK:*COMMUNICATION-STYLE*,
which specifies the mechanism by which Lisp reads and processes
protocol messages from Emacs. The choice of communication style has a
global influence on SLIME’s operation.
The available communication styles are:
NILThis style simply loops reading input from the communication socket and serves SLIME protocol events as they arise. The simplicity means that the Lisp cannot do any other processing while under SLIME’s control.
:FD-HANDLERThis style uses the classical Unix-style “select()-loop.”
Swank registers the communication socket with an event-dispatching
framework (such as SERVE-EVENT in CMUCL and
SBCL) and receives a callback when data is available. In
this style requests from Emacs are only detected and processed when
Lisp enters the event-loop. This style is simple and predictable.
:SIGIOThis style uses signal-driven I/O with a SIGIO signal
handler. Lisp receives requests from Emacs along with a signal,
causing it to interrupt whatever it is doing to serve the
request. This style has the advantage of responsiveness, since Emacs
can perform operations in Lisp even while it is busy doing other
things. It also allows Emacs to issue requests concurrently, e.g. to
send one long-running request (like compilation) and then interrupt
that with several short requests before it completes. The
disadvantages are that it may conflict with other uses of SIGIO
by Lisp code, and it may cause untold havoc by interrupting Lisp at an
awkward moment.
:SPAWNThis style uses multiprocessing support in the Lisp system to execute
each request in a separate thread. This style has similar properties
to :SIGIO, but it does not use signals and all requests issued
by Emacs can be executed in parallel.
The default request handling style is chosen according to the
capabilities of your Lisp system. The general order of preference is
:SPAWN, then :SIGIO, then :FD-HANDLER, with
NIL as a last resort. You can check the default style by
calling SWANK-BACKEND::PREFERRED-COMMUNICATION-STYLE. You can
also override the default by setting
SWANK:*COMMUNICATION-STYLE* in your Swank init file.
Previous: Communication style, Up: Lisp-side (Swank) [Contents][Index]
These Lisp variables can be configured via your ~/.swank.lisp file:
SWANK:*CONFIGURE-EMACS-INDENTATION*This variable controls whether indentation styles for
&body-arguments in macros are discovered and sent to Emacs. It
is enabled by default.
SWANK:*GLOBALLY-REDIRECT-IO*When T this causes the standard streams (*standard-output*,
etc) to be globally redirected to the REPL in Emacs.
When :STARTED-FROM-EMACS (default) redirects the output when
the lisp is launched from emacs (i.e. M-x slime), but not
from M-x slime-connect.
When NIL these streams are only temporarily redirected
to Emacs using dynamic bindings while handling requests. Note that
*standard-input* is currently never globally redirected into
Emacs, because it can interact badly with the Lisp’s native REPL by
having it try to read from the Emacs one.
SWANK:*GLOBAL-DEBUGGER*When true (the default) this causes *DEBUGGER-HOOK* to be
globally set to SWANK:SWANK-DEBUGGER-HOOK and thus for SLIME
to handle all debugging in the Lisp image. This is for debugging
multithreaded and callback-driven applications.
SWANK:*SLDB-QUIT-RESTART*This variable names the restart that is invoked when pressing q
(see sldb-quit) in SLDB. For SLIME evaluation requests this
is unconditionally bound to a restart that returns to a safe
point. This variable is supposed to customize what q does if an
application’s thread lands into the debugger (see
SWANK:*GLOBAL-DEBUGGER*).
(setf swank:*sldb-quit-restart* 'sb-thread:terminate-thread)
SWANK:*BACKTRACE-PRINTER-BINDINGS*SWANK:*MACROEXPAND-PRINTER-BINDINGS*SWANK:*SLDB-PRINTER-BINDINGS*SWANK:*SWANK-PPRINT-BINDINGS*These variables can be used to customize the printer in various situations. The values of the variables are association lists of printer variable names with the corresponding value. E.g., to enable the pretty printer for formatting backtraces in SLDB, you can use:
(push '(*print-pretty* . t) swank:*sldb-printer-bindings*).
SWANK:*LOG-EVENTS*Setting this variable to t causes all protocol messages
exchanged with Emacs to be printed to *TERMINAL-IO*. This is
useful for low-level debugging and for observing how SLIME works
“on the wire.” The output of *TERMINAL-IO* can be found in
your Lisp system’s own listener, usually in the buffer
*inferior-lisp*.
Next: Contributed Packages, Previous: Customization, Up: SLIME [Contents][Index]
Next: Globally redirecting all IO to the REPL, Up: Tips and Tricks [Contents][Index]
One of the advantages of the way SLIME is implemented is that we can easily run the Emacs side (slime.el) on one machine and the lisp backend (swank) on another. The basic idea is to start up lisp on the remote machine, load swank and wait for incoming SLIME connections. On the local machine we start up emacs and tell SLIME to connect to the remote machine. The details are a bit messier but the underlying idea is that simple.
Next: Setting up Emacs, Up: Connecting to a remote lisp [Contents][Index]
When you want to load swank without going through the normal, Emacs based, process just load the swank-loader.lisp file. Just execute
(load "/path/to/swank-loader.lisp") (swank-loader:init)
inside a running lisp image2. Now all we need to do is startup our swank server. The first example assumes we’re using the default settings.
(swank:create-server)
Since we’re going to be tunneling our connection via ssh3 and we’ll only have one port open we want to tell swank to not use an extra connection for output (this is actually the default in current SLIME):
(setf swank:*use-dedicated-output-stream* nil)
If you need to do anything particular
(like be able to reconnect to swank after you’re done), look into
swank:create-server’s other arguments. Some of these arguments
are
:PORTPort number for the server to listen on (default: 4005).
:STYLESee See Communication style.
:DONT-CLOSEBoolean indicating if the server will continue to accept connections
after the first one (default: NIL). For “long-running” lisp processes
to which you want to be able to connect from time to time,
specify :dont-close t
So the more complete example will be
(swank:create-server :port 4005 :dont-close t)
On the emacs side you will use something like
(setq slime-net-coding-system 'utf-8-unix) (slime-connect "localhost" 4005))
to connect to this lisp image from the same machine.
Next: Setting up pathname translations, Previous: Setting up the lisp image, Up: Connecting to a remote lisp [Contents][Index]
Now we need to create the tunnel between the local machine and the remote machine.
ssh -L4005:localhost:4005 username@remote.example.com
That ssh invocation creates an ssh tunnel between the port 4005 on our local machine and the port 4005 on the remote machine4.
Finally we can start SLIME:
M-x slime-connect RET RET
The RET RET sequence just means that we want to use the default
host (localhost) and the default port (4005). Even
though we’re connecting to a remote machine the ssh tunnel fools Emacs
into thinking it’s actually localhost.
Previous: Setting up Emacs, Up: Connecting to a remote lisp [Contents][Index]
One of the main problems with running swank remotely is that Emacs
assumes the files can be found using normal filenames. if we want
things like slime-compile-and-load-file (C-c C-k) and
slime-edit-definition (M-.) to work correctly we need to
find a way to let our local Emacs refer to remote files.
There are, mainly, two ways to do this. The first is to mount, using
NFS or similar, the remote machine’s hard disk on the local machine’s
file system in such a fashion that a filename like
/opt/project/source.lisp refers to the same file on both
machines. Unfortunately NFS is usually slow, often buggy, and not
always feasible, fortunately we have an ssh connection and Emacs’
tramp-mode can do the rest.
(See See (tramp)TRAMP User Manual.)
What we do is teach Emacs how to take a filename on the remote machine
and translate it into something that tramp can understand and access
(and vice versa). Assuming the remote machine’s host name is
remote.example.com, cl:machine-instance returns
“remote” and we login as the user “user” we can use slime-tramp
contrib to setup the proper translations by simply doing:
(add-to-list 'slime-filename-translations
(slime-create-filename-translator
:machine-instance "remote"
:remote-host "remote.example.com"
:username "user"))
Next: Connecting to SLIME automatically, Previous: Connecting to a remote lisp, Up: Tips and Tricks [Contents][Index]
When connecting via M-x slime-connect SLIME does
not change *standard-output* and friends outside of the
REPL. If you have any other threads which call format,
write-string, etc. that output will be seen only in
the *inferior-lisp* buffer or on the terminal, more often than
not this is inconvenient. So, if you want code such as this:
(run-in-new-thread
(lambda ()
(write-line "In some random thread.~%" *standard-output*)))
to send its output to SLIME’s repl buffer, as opposed to
*inferior-lisp*, set swank:*globally-redirect-io* to T
in ~/.swank.lisp
But when started using M-x slime the streams are redirected by default.
Previous: Globally redirecting all IO to the REPL, Up: Tips and Tricks [Contents][Index]
To make SLIME connect to your lisp whenever you open a lisp file just add this to your .emacs:
(add-hook 'slime-mode-hook
(lambda ()
(unless (slime-connected-p)
(save-excursion (slime)))))
Next: Credits, Previous: Tips and Tricks, Up: SLIME [Contents][Index]
In version 2.1 we moved some functionality to separate packages. This chapter tells you how to load contrib modules and describes what the particular packages do.
inferior-slime-modeslime-autodoc-modeslime-sprofslime-fancyNext: REPL: the “top level”, Up: Contributed Packages [Contents][Index]
Contrib packages aren’t loaded by default. You have to modify your
setup a bit so that Emacs knows where to find them and which of them
to load. Generally, you set the variable slime-contribs with
the list of package-names that you want to use. Its default value
is slime-fancy which loads almost everything. For example, a
setup to load the slime-scratch
and slime-editing-commands packages looks like:
;; Setup load-path and autoloads (add-to-list 'load-path "~/dir/to/cloned/slime") (require 'slime-autoloads) ;; Set your lisp system and some contribs (setq inferior-lisp-program "/opt/sbcl/bin/sbcl") (setq slime-contribs '(slime-scratch slime-editing-commands))
After starting SLIME, the commands of both packages should be available.
We recommend that you setup contribs before starting SLIME via
M-x slime, but if you want to enable more contribs after
you do that, you can set the slime-contribs variable to another
value and call M-x slime-setup. Note this though:
slime-setup step for each of them.
Short of restarting Emacs, a reasonable way of unloading contribs is
by calling an Emacs Lisp function whose name is obtained by
adding -unload to the contrib’s name, for every contrib you
wish to unload. So, to remove slime-repl, you must call
slime-repl-unload. Because the unload function will only, if
ever, unload the Emacs Lisp side of the contrib, you may also need to
restart your lisps.
Next: Multiple REPLs, Previous: Loading Contrib Packages, Up: Contributed Packages [Contents][Index]
SLIME uses a custom Read-Eval-Print Loop (REPL, also known as a
“top level”, or listener). The REPL user-interface is written in
Emacs Lisp, which gives more Emacs-integration than the traditional
comint-based Lisp interaction:
To load the REPL use (add-to-list 'slime-contribs 'slime-repl) in your
.emacs.
Select the output buffer, preferably in a different window.
Insert a call to the function defined around point into the REPL.
Inserts the last expression to the REPL and evaluates it there. Switches to the current package of the source buffer for the duration. If used with a prefix argument, doesn’t switch back afterwards.
Next: Input navigation, Up: REPL: the “top level” [Contents][Index]
Evaluate the current input in Lisp if it is complete. If incomplete, open a new line and indent. If a prefix argument is given then the input is evaluated without checking for completeness.
Close any unmatched parenthesis and then evaluate the current input in Lisp. Also bound to M-RET.
Indent the current line and perform symbol completion.
Open and indent a new line.
Go to the beginning of the line, but stop at the REPL prompt.
Interrupt the Lisp process with SIGINT.
Clear the entire buffer, leaving only a prompt.
Remove the output and result of the previous expression from the buffer.
Previous: Input navigation, Up: REPL: the “top level” [Contents][Index]
“Shortcuts” are a special set of REPL commands that are invoked by name. To invoke a shortcut you first press , (comma) at the REPL prompt and then enter the shortcut’s name when prompted.
Shortcuts deal with things like switching between directories and
compiling and loading Lisp systems. The set of shortcuts is listed
below, and you can also use the help
shortcut to list them interactively.
Change the current directory.
Change the current package.
Compile (if necessary) and load a lisp file.
Define a new global, special, variable.
Disconnect all connections.
Display the help.
Pop the current directory.
Pop the top of the package stack.
Push a new directory onto the directory stack.
Push a package onto the package stack.
Show the current directory.
Quit the current Lisp.
Resend the last form.
Restart *inferior-lisp* and reconnect SLIME.
Quit all Lisps and close all SLIME buffers.
Next: inferior-slime-mode, Previous: REPL: the “top level”, Up: Contributed Packages [Contents][Index]
The slime-mrepl package adds support for multiple listener
buffers. The command M-x slime-new-mrepl creates a new
buffer. In a multi-threaded Lisp, each listener is associated with a
separate thread. In a single-threaded Lisp it’s also possible to
create multiple listener buffers but the commands are executed
sequentially by the same process.
Next: Compound Completion, Previous: Multiple REPLs, Up: Contributed Packages [Contents][Index]
inferior-slime-modeThe inferior-slime-mode is a minor mode is intended to use with
the *inferior-lisp* lisp buffer. It provides some of the
SLIME commands, like symbol completion and documentation lookup. It
also tracks the current directory of the Lisp process. To install it,
add something like this to user .emacs:
(add-to-list 'slime-contribs 'inferior-slime)
Turns inferior-slime-mode on or off.
The variable inferior-slime-mode-map contains the extra
keybindings.
Next: Fuzzy Completion, Previous: inferior-slime-mode, Up: Contributed Packages [Contents][Index]
The package slime-c-p-c provides a different symbol completion
algorithm, which performs completion “in parallel” over the
hyphen-delimited sub-words of a symbol name.
5
Formally this means that “a-b-c” can complete to any symbol
matching the regular expression “^a.*-b.*-c.*” (where “dot”
matches anything but a hyphen). Examples give a more intuitive
feeling:
m-v-b completes to multiple-value-bind.
w-open is ambiguous: it completes to either
with-open-file or with-open-stream. The symbol is
expanded to the longest common completion (with-open-) and the
point is placed at the first point of ambiguity, which in this case is
the end.
w--stream completes to with-open-stream.
The variable slime-c-p-c-unambiguous-prefix-p specifies where
point should be placed after completion. E.g. the possible
completions for f-o are finish-output and
force-output. By the default point is moved after the
f, because that is the unambiguous prefix. If
slime-c-p-c-unambiguous-prefix-p is nil, point moves to
the end of the inserted text, after the o in this case.
In addition, slime-c-p-c provides completion for character names
(mostly useful for Unicode-aware implementations):
CL-USER> #\Sp<TAB>
Here SLIME will usually complete the character to #\Space, but
in a Unicode-aware implementation, this might provide the following
completions:
Space Space Sparkle Spherical_Angle Spherical_Angle_Opening_Left Spherical_Angle_Opening_Up
The package slime-c-p-c also provides context-sensitive
completion for keywords. Example:
CL-USER> (find 1 '(1 2 3) :s<TAB>
Here SLIME will complete :start, rather than suggesting all
ever-interned keywords starting with :s.
Looks up and inserts into the current buffer the argument list for the
function at point, if there is one. More generally, the command
completes an incomplete form with a template for the missing arguments.
There is special code for discovering extra keywords of generic
functions and for handling make-instance,
defmethod, and many other functions. Examples:
(subseq "abc" <C-c C-s>
--inserts--> start [end])
(find 17 <C-c C-s>
--inserts--> sequence :from-end from-end :test test
:test-not test-not :start start :end end
:key key)
(find 17 '(17 18 19) :test #'= <C-c C-s>
--inserts--> :from-end from-end
:test-not test-not :start start :end end
:key key)
(defclass foo () ((bar :initarg :bar)))
(defmethod print-object <C-c C-s>
--inserts--> (object stream)
body...)
(defmethod initialize-instance :after ((object foo) &key blub))
(make-instance 'foo <C-c C-s>
--inserts--> :bar bar :blub blub initargs...)
Next: slime-autodoc-mode, Previous: Compound Completion, Up: Contributed Packages [Contents][Index]
The package slime-fuzzy implements yet another symbol
completion heuristic.
Presents a list of likely completions to choose from for an
abbreviation at point. If you set the
variable slime-complete-symbol-function to this command, fuzzy
completion will also be used for M-TAB.
It attempts to complete a symbol all at once, instead of in pieces.
For example, “mvb” will find “multiple-value-bind” and
“norm-df” will find
“least-positive-normalized-double-float”.
The algorithm tries to expand every character in various ways and rates the list of possible completions with the following heuristic.
Letters are given scores based on their position in the string. Letters at the beginning of a string or after a prefix letter at the beginning of a string are scored highest. Letters after a word separator such as #\- are scored next highest. Letters at the end of a string or before a suffix letter at the end of a string are scored medium, and letters anywhere else are scored low.
If a letter is directly after another matched letter, and its intrinsic value in that position is less than a percentage of the previous letter’s value, it will use that percentage instead.
Finally, a small scaling factor is applied to favor shorter matches, all other things being equal.
In case a symbol is accessible via several packages, duplicate symbol
filter specified via *fuzzy-duplicate-symbol-filter* swank
variable is applied. :nearest-package value specifies that only
symbols in the package with highest score should be kept.
:home-package specifies that only the match that represents the home
package of the symbol is used, and :all value specifies that
duplicate symbol filter mode should be turned off.
To specify a custom filter, set *fuzzy-duplicate-symbol-filter*
to a function accepting three arguments: the name of package being
examined, the list of names of all packages being examined with
packages with highest matching score listed first and an equal
hash-table that is shared between calls to the function and can be
used for deduplication purposes. The function should return a
deduplication filter function which accepts a symbol and returns true
if the symbol should be kept.
For example, the effect of :nearest-package can be also achieved
by specifying the following custom filter in ~/.swank.lisp:
(setf *fuzzy-duplicate-symbol-filter*
(lambda (cur-package all-packages dedup-table)
(declare (ignore cur-package all-packages))
(lambda (symbol)
(unless (gethash (symbol-name symbol) dedup-table)
(setf (gethash (symbol-name symbol) dedup-table) t)))))
And instead of :home-package, the following can be used:
(setf *fuzzy-duplicate-symbol-filter*
(lambda (cur-package all-packages dedup-table)
(declare (ignore dedup-table))
(let ((packages (mapcar #'find-package
(remove cur-package all-packages))))
(lambda (symbol)
(not (member (symbol-package symbol) packages))))))
Next: ASDF, Previous: Fuzzy Completion, Up: Contributed Packages [Contents][Index]
slime-autodoc-modeAutodoc mode is an additional minor-mode for automatically showing
information about symbols near the point. For function names the
argument list is displayed, and for global variables, the value.
Autodoc is implemented by means of eldoc-mode of Emacs.
The mode can be enabled by default in your ~/.emacs:
(add-to-list 'slime-contribs 'slime-autodoc)
Show the argument list of the function NAME.
Toggles autodoc-mode on or off according to the argument, and toggles the mode when invoked without argument.
Like slime-autodoc, but when called twice, or after slime-autodoc was already automatically called, display multiline arglist.
If the variable slime-use-autodoc-mode is set (default), Emacs
starts a timer, otherwise the information is only displayed after
pressing SPC.
If slime-autodoc-use-multiline-p is set to non-nil,
allow long autodoc messages to resize echo area display.
slime-autodoc-mode-string is a string that will be displayed in
the mode line when autodoc-mode is enabled, or nil, if you prefer no
indication. You can customize this variable.
Next: Banner, Previous: slime-autodoc-mode, Up: Contributed Packages [Contents][Index]
ASDF is a popular “system construction tool”. The package
slime-asdf provides some commands to load and compile such
systems from Emacs. ASDF itself is not included with
SLIME; you have to load that yourself into your Lisp. In
particular, you must load ASDF before you connect, otherwise
you will get errors about missing symbols.
Compile and load an ASDF system. The default system name is taken from the first file matching *.asd in the current directory.
Recompile and load an ASDF system without recompiling its dependencies.
Open all files in a system, optionally load it if LOAD is non-nil.
Browse files in a system using Dired.
Delete FASLs produced by compiling a system.
Run rgrep on the base directory of an ASDF system.
Run isearch-forward on the files of an ASDF system.
Run query-replace on an ASDF system.
The package also installs some new REPL shortcuts (see Shortcuts):
Compile (as needed) and load an ASDF system.
Recompile and load an ASDF system.
Compile (but not load) an ASDF system.
Recompile (but not load) an ASDF system.
Recompile and load an ASDF system.
Open all files in a system.
Browse files in a system using Dired.
Delete FASLs produced by compiling a system.
Next: Editing Commands, Previous: ASDF, Up: Contributed Packages [Contents][Index]
The package slime-banner installs a window header line (
See (elisp)Header Lines.) in the REPL buffer. It also runs an
animation at startup.
By setting the variable slime-startup-animation to nil you can
disable the animation respectively with the
variable slime-header-line-p the header line.
Next: Fancy Inspector, Previous: Banner, Up: Contributed Packages [Contents][Index]
The package slime-editing-commands provides some commands to
edit Lisp expressions.
Re-indents the current defun, or refills the current paragraph.
If point is inside a comment block, the text around point will be
treated as a paragraph and will be filled with fill-paragraph.
Otherwise, it will be treated as Lisp code, and the current defun
will be reindented. If the current defun has unbalanced parens,
an attempt will be made to fix it before reindenting.
Balance parentheses of open s-expressions at point. Insert enough right parentheses to balance unmatched left parentheses. Delete extra left parentheses. Reformat trailing parentheses Lisp-stylishly.
If REGION is true, operate on the region. Otherwise operate on the top-level sexp before point.
Insert a set of balanced comments around the s-expression containing the point. If this command is invoked repeatedly (without any other command occurring between invocations), the comment progressively moves outward over enclosing expressions. If invoked with a positive prefix argument, the s-expression arg expressions out is enclosed in a set of balanced comments.
Next: Presentations, Previous: Editing Commands, Up: Contributed Packages [Contents][Index]
An alternative to default inspector is provided by the package ‘slime-fancy-inspector’. This inspector knows a lot about CLOS objects and methods. It provides many “actions” that can be selected to invoke Lisp code on the inspected object. For example, to present a generic function the inspector shows the documentation in plain text and presents each method with both a hyperlink to inspect the method object and a “remove method” action that you can invoke interactively. The key-bindings are the same as for the basic inspector (see Inspector commands).
Next: Typeout frames, Previous: Fancy Inspector, Up: Contributed Packages [Contents][Index]
A “presentation”6 in SLIME is a region of text associated with a Lisp object. Right-clicking on the text brings up a menu with operations for the particular object. Some operations, like inspecting, are available for all objects, but the object may also have specialized operations. For instance, pathnames have a dired operation.
More importantly, it is possible to cut and paste presentations (i.e., Lisp objects, not just their printed presentation), using all standard Emacs commands. This way it is possible to cut and paste the results of previous computations in the REPL. This is of particular importance for unreadable objects.
The package slime-presentations installs presentations in the
REPL, i.e. the results of evaluation commands become presentations. In
this way, presentations generalize the use of the standard Common Lisp
REPL history variables *, **, ***. Example:
CL-USER> (find-class 'standard-class) #<STANDARD-CLASS STANDARD-CLASS> CL-USER>
Presentations appear in red color in the buffer. (In this manual, we indicate the presentations like this.) Using standard Emacs commands, the presentation can be copied to a new input in the REPL:
CL-USER> (eql '#<STANDARD-CLASS STANDARD-CLASS>
'#<STANDARD-CLASS STANDARD-CLASS>)
T
Note that standard evaluation and quoting rules still apply. So if a presentation is a list, it needs to be quoted in an evaluated context to avoid treating it as a function call:
CL-USER> (list (find-class 'standard-class) 2 3 4) (#<STANDARD-CLASS STANDARD-CLASS> 2 3 4) CL-USER> (#<STANDARD-CLASS STANDARD-CLASS> 2 3 4) ; Funcall of #<STANDARD-CLASS STANDARD-CLASS> which is a non-function. ; Evaluation aborted. CL-USER> '(#<STANDARD-CLASS STANDARD-CLASS> 2 3 4) (#<STANDARD-CLASS STANDARD-CLASS> 2 3 4)
When you copy an incomplete presentation or edit the text within a presentation, the presentation changes to plain text, losing the association with a Lisp object. In the buffer, this is indicated by changing the color of the text from red to black. This can be undone.
Presentations are also available in the inspector (all inspectable parts
are presentations) and the debugger (all local variables are
presentations). This makes it possible to evaluate expressions in the
REPL using objects that appear in local variables of some active
debugger frame; this can be more convenient than using M-x
sldb-eval-in-frame. Warning: The presentations that stem from
the inspector and debugger are only valid as long as the corresponding
buffers are open. Using them later can cause errors or confusing
behavior.
For some Lisp implementations you can also install the package
slime-presentation-streams, which enables presentations on the
Lisp *standard-output* stream and similar streams. This means
that not only results
of computations, but also some objects that are printed to the standard
output (as a side-effect of the computation) are associated with
presentations. Currently, all unreadable objects
and pathnames get printed as presentations.
CL-USER> (describe (find-class 'standard-object))
#<STANDARD-CLASS STANDARD-OBJECT> is an instance of
#<STANDARD-CLASS STANDARD-CLASS>:
The following slots have :INSTANCE allocation:
PLIST NIL
FLAGS 1
DIRECT-METHODS ((#<STANDARD-METHOD
SWANK::ALL-SLOTS-FOR-INSPECTOR
(STANDARD-OBJECT T)>
...
Again, this makes it possible to inspect and copy-paste these objects.
In addition to the standard Emacs commands, there are several keyboard commands, a menu-bar menu, and a context menu to operate on presentations. We describe the keyboard commands below; they are also shown in the menu-bar menu.
If point is within a presentation, move point to the beginning of the presentation and mark to the end of the presentation. This makes it possible to copy the presentation.
If point is within a presentation, copy the surrounding presentation to the kill ring.
If point is within a presentation, copy the surrounding presentation to the REPL.
If point is within a presentation, describe the associated object.
If point is within a presentation, inspect the associated object with the SLIME inspector.
Move point to the next presentation in the buffer.
Move point to the previous presentation in the buffer.
Similar operations are also possible from the context menu of every
presentation. Using mouse-3 on a presentation, the context menu
opens and offers various commands. For some objects, specialized
commands are also offered. Users can define additional specialized
commands by defining a method for
swank::menu-choices-for-presentation.
Warning: On Lisp implementations without weak hash tables,
all objects associated with presentations are protected from garbage
collection. If your Lisp image grows too large because of that,
use C-c C-v M-o (slime-clear-presentations) to remove these
associations. You can also use the command C-c M-o
(slime-repl-clear-buffer), which both clears the REPL buffer and
removes all associations of objects with presentations.
Warning: Presentations can confuse new users.
CL-USER> (cons 1 2) (1 . 2) CL-USER> (eq '(1 . 2) '(1 . 2)) T
One could have expected NIL here, because it looks like two
fresh cons cells are compared regarding object identity.
However, in the example the presentation (1 . 2) was copied twice
to the REPL. Thus EQ is really invoked with the same object,
namely the cons cell that was returned by the first form entered in the
REPL.
Next: TRAMP, Previous: Presentations, Up: Contributed Packages [Contents][Index]
A “typeout frame” is a special Emacs frame which is used instead of the echo area (minibuffer) to display messages from SLIME commands. This is an optional feature. The advantage of a typeout frame over the echo area is that it can hold more text, it can be scrolled, and its contents don’t disappear when you press a key. All potentially long messages are sent to the typeout frame, such as argument lists, macro expansions, and so on.
Ensure that a typeout frame exists, creating one if necessary.
If the typeout frame is closed then the echo area will be used again as usual.
To have a typeout frame created automatically at startup you should
load the slime-typeout-frame package. (see Loading Contrib Packages.)
The variable slime-typeout-frame-properties specifies the
height and possibly other properties of the frame. Its value is
passed to make-frame. (See (elisp)Creating Frames.)
Next: Documentation Links, Previous: Typeout frames, Up: Contributed Packages [Contents][Index]
The package slime-tramp provides some functions to set up
filename translations for TRAMP. (see Setting up pathname translations)
Next: Xref and Class Browser, Previous: TRAMP, Up: Contributed Packages [Contents][Index]
For certain error messages, SBCL includes references to the ANSI
Standard or the SBCL User Manual. The slime-references package
turns those references into clickable links. This makes finding the
referenced section of the HyperSpec much easier.
Next: Highlight Edits, Previous: Documentation Links, Up: Contributed Packages [Contents][Index]
A rudimentary class browser is provided by
the slime-xref-browser package.
This command asks for a class name and displays inheritance tree of for the class.
This command prompts for a symbol and the kind of cross reference, e.g. callers. The cross reference tree rooted at the symbol is then then displayed.
Next: Scratch Buffer, Previous: Xref and Class Browser, Up: Contributed Packages [Contents][Index]
slime-highlight-edits is a minor mode to highlight those
regions in a Lisp source file which are modified. This is useful to
quickly find those functions which need to be recompiled (with
C-c C-c)
Turns slime-highlight-edits-mode on or off.
Next: SLIME Trace Dialog, Previous: Highlight Edits, Up: Contributed Packages [Contents][Index]
The SLIME scratch buffer, in contrib package slime-scratch,
imitates Emacs’ usual *scratch* buffer.
If slime-scratch-file is set, it is used to back the scratch
buffer, making it persistent. The buffer is like any other Lisp
buffer, except for the command bound to C-j.
Evaluate the expression sexp before point and insert print value into the current buffer.
Create a *slime-scratch* buffer. In this buffer you can enter Lisp expressions and evaluate them with C-j, like in Emacs’s *scratch* buffer.
Next: slime-sprof, Previous: Scratch Buffer, Up: Contributed Packages [Contents][Index]
The SLIME Trace Dialog, in package slime-trace-dialog, is a
tracing facility, similar to Common Lisp’s trace, but
interactive rather than purely textual. It is an Emacs 24-only
contrib.
You use it just like you would regular trace: after tracing a
function, calling it causes interesting information about that
particular call to be reported.
However, instead of printing the trace results to the
the *trace-output* stream (usually the REPL), the SLIME
Trace Dialog collects and stores them in your lisp environment until,
on user’s request, they are fetched into Emacs and displayed in a
dialog-like interactive view.
After starting up SLIME, SLIME’s Trace Dialog installs
a Trace menu in the menu-bar of any slime-mode buffer and
adds two new commands, with respective key-bindings:
If point is on a symbol name, toggle tracing of its function definition. If point is not on a symbol, prompt user for a function.
With a C-u prefix argument, and if your lisp implementation allows it, attempt to decipher lambdas, methods and other complicated function signatures.
The function is traced for the SLIME Trace Dialog only, i.e. it is
not found in the list returned by Common Lisp’s trace.
Pop to the interactive SLIME Trace Dialog buffer associated with the current connection (see Multiple connections).
Consider the (useless) program:
(defun foo (n) (if (plusp n) (* n (bar (1- n))) 1)) (defun bar (n) (if (plusp n) (* n (foo (1- n))) 1))
After tracing both foo and bar with C-c M-t,
calling call (foo 2) and moving to the trace dialog with
C-c T, we are presented with this buffer.
Traced specs (2) [refresh]
[untrace all]
[untrace] common-lisp-user::bar
[untrace] common-lisp-user::foo
Trace collection status (3/3) [refresh]
[clear]
0 - common-lisp-user::foo
| > 2
| < 2
1 `--- common-lisp-user::bar
| > 1
| < 1
2 `-- common-lisp-user::foo
> 0
< 1
The dialog is divided into sections displaying the functions already traced, the trace collection progress and the actual trace tree that follow your program’s logic. The most important key-bindings in this buffer are:
Update information on the trace collection and traced specs.
Fetch the next batch of outstanding (not fetched yet) traces. With a C-u prefix argument, repeat until no more outstanding traces.
Prompt for confirmation, then clear all traces, both fetched and outstanding.
The arguments and return values below each entry are interactive
buttons. Clicking them opens the inspector
(see Inspector commands). Invoking M-RET
(slime-trace-dialog-copy-down-to-repl) returns them to the REPL
for manipulation (see REPL: the “top level”). The number left of each entry
indicates its absolute position in the calling order, which might
differ from display order in case multiple threads call the same
traced function.
slime-trace-dialog-hide-details-mode hides arguments and return
values so you can concentrate on the calling logic. Additionally,
slime-trace-dialog-autofollow-mode will automatically
display additional detail about an entry when the cursor moves over
it.
Next: SLIME Enhanced M-., Previous: SLIME Trace Dialog, Up: Contributed Packages [Contents][Index]
slime-sprofslime-sprof is a package for integrating SBCL’s statistical profiler, sb-sprof.
The variable slime-sprof-exclude-swank controls whether to
display swank functions. The default value is NIL.
Start profiling.
Stop profiling.
Report results of the profiling.
The following keys are defined in slime-sprof-browser mode:
Expand / collapse function details (callers, calls to)
View function sources.
Disassemble function.
Toggle exclusion of swank functions from the report.
Next: Meta package: slime-fancy, Previous: slime-sprof, Up: Contributed Packages [Contents][Index]
slime-mdot-fu enables meta-point to jump to local variables
bound with let and let*, in addition to function bindings
declared with flet and labels, via
slime-edit-local-definition.
Next: Quicklisp, Previous: SLIME Enhanced M-., Up: Contributed Packages [Contents][Index]
slime-fancyslime-fancy is a meta package which loads a combination of the
most popular packages. It is the default value of
slime-contribs, thus it is loaded by default.
Previous: Meta package: slime-fancy, Up: Contributed Packages [Contents][Index]
The package slime-quicklisp adds support for loading Quicklisp
systems in the REPL buffer. In order for this to work, Quicklisp
should have already been loaded in the Lisp implementation. Refer
to https://www.quicklisp.org/ for Quicklisp installation
details.
The package installs the following REPL shortcuts (see Shortcuts):
Load a Quicklisp system.
Next: Key (Character) Index, Previous: Contributed Packages, Up: SLIME [Contents][Index]
The soppy ending...
SLIME is an Extension of SLIM by Eric Marsden. At the time of writing, the authors and code-contributors of SLIME are:
| Helmut Eller | Tobias C. Rittweiler | Stas Boukarev |
| Luke Gorrie | Matthias Koeppe | LuÃs Oliveira |
| Nikodemus Siivola | Marco Baringer | João Távora |
| Alan Ruttenberg | Mark Evenson | Henry Harrington |
| Christophe Rhodes | Edi Weitz | Martin Simmons |
| Juho Snellman | Attila Lendvai | Peter Seibel |
| Geo Carncross | Daniel Kochmanski | Douglas Crosher |
| Chris Schafmeister | Gábor Melis | Daniel Barlow |
| Wolfgang Jenkner | LuÃs Borges de Oliveira | Jan Moringen |
| Stelian Ionescu | Michael Weber | Didier Verna |
| Lawrence Mitchell | Anton Kovalenko | Terje Norderhaug |
| Mark | Brian Downing | Bill Clementson |
| Andras Simon | Adlai Chandrasekhar | Zach Beane |
| Tarn W. Burton | Ivan Shvedunov | Gabor Melis |
| Francois-Rene Rideau | Espen Wiborg | Christian Schafmeister |
| Bike | António Menezes Leitão | Alexander Artemenko |
| Utz-Uwe Haus | Thomas Schilling | Thomas F. Burdick |
| Takehiko Abe | Sébastien Villemot | Richard M Kreuter |
| Raymond Toy | Matthew Danish | Mark Harig |
| James Bielman | Harald Hanche-Olsen | Ed Langley |
| Ariel Badichi | Andreas Fuchs | Willem Broekema |
| Taylor R. Campbell | Steve Purcell | Phil Hargett |
| Paulo Madeira | Nicolas Martyanoff | Lars Magne Ingebrigtsen |
| Kris Katterjohn | John Paul Wallington | Joerg Hoehle |
| Douglas Katzman | David Reitter | Charles Zhang |
| Bryan O’Connor | Bruno Cichon | Bart Botta |
| Alan Shutko | Ursa americanus kermodei | Travis Cross |
| Tobias Rittweiler | Tiago Maduro-Dias | Stefan Kamphausen |
| Sean O’Rourke | Robert Lehr | Robert E. Brown |
| Robert Brown | Renke Christian von Seggern | Philipp Stephani |
| Philipp Marek | Peter S. Housel | Nathan Trapuzzano |
| Nathan Bird | Mike Appleby | Matteo Landi |
| Knut Olav Bøhmer | Jouni K Seppanen | Jon Oddie |
| Ivan Toshkov | Ian Eslick | Geoff Wozniak |
| Gary King | Fice T | Eric Blood |
| Eduardo Muñoz | Christophe Junke | Christian Lynbech |
| Chris Capel | Bjørn Nordbø | Anton Vodonosov |
| Alexey Dejneka | Alan Caulkins | Zachary Beane |
| Yu-Chiang Hsu | Yaroslav Kavenchuk | Yan |
| YOKOTA Yuki | Wolfgang Mederle | Wojciech Kaczmarek |
| William Bland | Wesley Harvey | Vitaly Mayatskikh |
| Tomas Zellerin | Tom Pierce | Tim Daly Jr. |
| Thomas Fitzsimmons | The Gendl Project | Syohei YOSHIDA |
| Sven Van Caekenberghe | Svein Ove Aas | Steve Smith |
| StanisBaw Halik | Sergey Kostyaev | Samuel Freilich |
| Russell Sim | Russell McManus | Russ Tyndall |
| Rui PatrocÃnio | Robert P. Goldman | Robert Macomber |
| Richard Garner | Reini Urban | R. Matthew Emerson |
| R Primus | Peter Feigl | Peter |
| Pawel Ostrowski | Paul Eggert | Paul Donnelly |
| Paul Collins | Patrick Poitras | Olof-Joachim Frahm |
| Neil Van Dyke | Nathan Ringo | NIIMI Satoshi |
| Mészáros Levente | Mikel Bancroft | Michał Herda |
| Michał "phoe" Herda | Michael White | Matthew Kennedy |
| Matthew D. Swank | Matt Pillsbury | Masayuki Onjo |
| Mark Wooding | Mark Karpov | Mark H. David |
| Marius Gerbershagen | Marco Monteiro | Lynn Quam |
| Levente Mészáros | Leo Liu | Lasse Rasinen |
| Kasper Gałkowski | Kasper | Karsten Poeck |
| Kai Kaminski | Julian Stecklina | Juergen Gmeiner |
| Jon Allen Boone | John Stracke | John Smith |
| Johan Bockgård | Joe Robertson | Jimmy Aguilar Mena |
| Jim Newton | Javier Olaechea | Jan Rychter |
| James McIlree | Jack Pugmire | Jacek Podkanski |
| Ivan Sokolov | Ivan Boldyrev | Ignas Mikalajunas |
| Hannu Koivisto | Grant Shangreaux | Graham Dobbins |
| Gleefre | Gerd Flaig | Gail Zacharias |
| Frederic Brunel | Eric Timmons | Eric Schulte |
| Dustin Long | Dmitry Igrishin | Deokhwan Kim |
| Denis Budyak | Daniel Koning | Daniel Kochmański |
| Dan Weinreb | Dan Pierson | Dacoda Strack |
| Cyrus Harmon | Christoph Keßler | Cecil Westerhof |
| Brian Mastenbrook | Brandon Bergren | Bozhidar Batsov |
| Bob Halley | Barry Fishman | B.Scott Michel |
| Angelo Rossi | Andrew Myers | Andre A. Gomes |
| Alexander Konstantinov | Aleksandar Bakic | Alain Picard |
| Al Hoang | Adam Bozanich |
... not counting the bundled code from hyperspec.el, CLOCC, and the CMU AI Repository.
Many people on the slime-devel mailing list have made non-code
contributions to SLIME. Life is hard though: you gotta send code to
get your name in the manual. :-)
We’re indebted to the good people of common-lisp.net for their
hosting and help, and for rescuing us from “Sourceforge hell.”
Implementors of the Lisps that we support have been a great help. We’d like to thank the CMUCL maintainers for their helpful answers, Craig Norvell and Kevin Layer at Franz providing Allegro CL licenses for SLIME development, and Peter Graves for his help to get SLIME running with ABCL.
Most of all we’re happy to be working with the Lisp implementors who’ve joined in the SLIME development: Dan Barlow and Christophe Rhodes of SBCL, Gary Byers of OpenMCL, and Martin Simmons of LispWorks. Thanks also to Alain Picard and Memetrics for funding Martin’s initial work on the LispWorks backend!
Next: Command and Function Index, Previous: Credits, Up: SLIME [Contents][Index]
| Jump to: | .
:
<
>
A B C D E G I L M N O P Q R S T V X |
|---|
| Jump to: | .
:
<
>
A B C D E G I L M N O P Q R S T V X |
|---|
Next: Variable and Concept Index, Previous: Key (Character) Index, Up: SLIME [Contents][Index]
| Jump to: | H I N S |
|---|
| Jump to: | H I N S |
|---|
Previous: Command and Function Index, Up: SLIME [Contents][Index]
| Jump to: | A C D I L M P S T U X |
|---|
| Jump to: | A C D I L M P S T U X |
|---|
Of course we made sure it was actually too slow before making the ugly optimization.
SLIME also provides an ASDF system definition which does the same thing
there is a way to connect without an ssh tunnel, but it has the side-effect of giving the entire world access to your lisp image, so we’re not going to talk about it
By
default swank listens for incoming connections on port 4005, had we
passed a :port parameter to swank:create-server we’d be
using that port number instead
This style of completion is modelled on completer.el by Chris McConnell. That package is bundled with ILISP.
Presentations are a feature originating
from the Lisp machines. It was possible to define present
methods specialized to various devices, e.g. to draw an object to
bitmapped screen or to write some text to a character stream.