Mercurial > emacs

@setfilename LNEWS

@section New Features in the Lisp Language

@end itemize
@itemize @bullet
@item
The new function @code{delete} is a traditional Lisp function.  It takes
two arguments, @var{elt} and @var{list}, and deletes from @var{list} any
elements that are equal to @var{elt}.  It uses the function @code{equal}
to compare elements with @var{elt}.

@item
The new function @code{member} is a traditional Lisp function.  It takes
two arguments, @var{elt} and @var{list}, and finds the first element of
@var{list} that is equal to @var{elt}.  It uses the function
@code{equal} to compare each list element with @var{elt}.

The value is a sublist of @var{list}, whose first element is the one
that was found.  If no matching element is found, the value is
@code{nil}.

@ignore @c Seems not to be true, from looking at the code.
@item
The function @code{equal} is now more robust: it does not crash due to
circular list structure.
@end ignore

@item
The new function @code{indirect-function} finds the effective function
definition of an object called as a function.  If the object is a
symbol, @code{indirect-function} looks in the function definition of the
symbol.  It keeps doing this until it finds something that is not a
symbol.

@item
There are new escape sequences for use in character and string
constants.  The escape sequence @samp{\a} is equivalent to @samp{\C-g},
the @sc{ASCII} @sc{BEL} character (code 7).  The escape sequence
@samp{\x} followed by a hexidecimal number represents the character
whose @sc{ASCII} code is that number.  There is no limit on the number
of digits in the hexidecimal value.

@item
The function @code{read} when reading from a buffer now does not skip a
terminator character that terminates a symbol.  It leaves that character
to be read (or just skipped, if it is whitespace) next time.

@item
When you use a function @var{function} as the input stream for
@code{read}, it is usually called with no arguments, and should return
the next character.  In Emacs 19, sometimes @var{function} is called
with one argument (always a character).  When that happens,
@var{function} should save the argument and arrange to return it when
called next time.

@item
@code{random} with integer argument @var{n} returns a random number
between 0 and @var{n}@minus{}1.

@item
The functions @code{documentation} and @code{documentation-property} now
take an additional optional argument which, if non-@code{nil}, says to
refrain from calling @code{substitute-command-keys}.  This way, you get
the exact text of the documentation string as written, without the usual
substitutions.  Make sure to call @code{substitute-command-keys}
yourself if you decide to display the string.

@ignore
@item
The new function @code{invocation-name} returns as a string the program
name that was used to run Emacs, with any directory names discarded.
@c ??? This hasn't been written yet. ???
@end ignore

@item
The new function @code{map-y-or-n-p} makes it convenient to ask a series
of similar questions.  The arguments are @var{prompter}, @var{actor},
@var{list}, and optional @var{help}.

The value of @var{list} is a list of objects, or a function of no
arguments to return either the next object or @code{nil} meaning there
are no more.

The argument @var{prompter} specifies how to ask each question.  If
@var{prompter} is a string, the question text is computed like this:

@example
(format @var{prompter} @var{object})
@end example

@noindent
where @var{object} is the next object to ask about.

If not a string, @var{prompter} should be a function of one argument
(the next object to ask about) and should return the question text.

The argument @var{actor} should be a function of one argument, which is
called with each object that the user says yes for.  Its argument is
always one object from @var{list}.

If @var{help} is given, it is a list @code{(@var{object} @var{objects}
@var{action})}, where @var{object} is a string containing a singular
noun that describes the objects conceptually being acted on;
@var{objects} is the corresponding plural noun and @var{action} is a
transitive verb describing @var{actor}.  The default is @code{("object"
"objects" "act on")}.

Each time a question is asked, the user may enter @kbd{y}, @kbd{Y}, or
@key{SPC} to act on that object; @kbd{n}, @kbd{N}, or @key{DEL} to skip
that object; @kbd{!} to act on all following objects; @key{ESC} or
@kbd{q} to exit (skip all following objects); @kbd{.} (period) to act on
the current object and then exit; or @kbd{C-h} to get help.

@code{map-y-or-n-p} returns the number of objects acted on.

@item
You can now ``set'' environment variables with the @code{setenv}
command.  This works by setting the variable @code{process-environment},
which @code{getenv} now examines in preference to the environment Emacs
received from its parent.
@end itemize

@section New Features for Loading Libraries

You can now arrange to run a hook if a particular Lisp library is
loaded.

The variable @code{after-load-alist} is an alist of expressions to be
evalled when particular files are loaded.  Each element looks like
@code{(@var{filename} @var{forms}@dots{})}.

When @code{load} is run and the file name argument equals
@var{filename}, the @var{forms} in the corresponding element are
executed at the end of loading.  @var{filename} must match exactly!
Normally @var{filename} is the name of a library, with no directory
specified, since that is how @code{load} is normally called.

An error in @var{forms} does not undo the load, but does prevent
execution of the rest of the @var{forms}.

The function @code{eval-after-load} provides a convenient way to add
entries to the alist.  Call it with two arguments, @var{file} and a
form to execute.

The function @code{autoload} now supports autoloading a keymap.
Use @code{keymap} as the fourth argument if the autoloaded function
will become a keymap when loaded.

There is a new feature for specifying which functions in a library should
be autoloaded by writing special ``magic'' comments in that library itself.

  Write @samp{;;;###autoload} on a line by itself before a function
definition before the real definition of the function, in its
autoloadable source file; then the command @kbd{M-x
update-file-autoloads} automatically puts the @code{autoload} call into
@file{loaddefs.el}.

  You can also put other kinds of forms into @file{loaddefs.el}, by
writing @samp{;;;###autoload} followed on the same line by the form.
@kbd{M-x update-file-autoloads} copies the form from that line.

@section Compilation Features

@itemize @bullet
@item
Inline functions.

You can define an @dfn{inline function} with @code{defsubst}.  Use
@code{defsubst} just like @code{defun}, and it defines a function which
you can call in all the usual ways.  Whenever the function thus defined
is used in compiled code, the compiler will open code it.

You can get somewhat the same effects with a macro, but a macro has the
limitation that you can use it only explicitly; a macro cannot be called
with @code{apply}, @code{mapcar} and so on.  Also, it takes some work to
convert an ordinary function into a macro.  To convert it into an inline
function, simply replace @code{defun} with @code{defsubst}.

Making a function inline makes explicit calls run faster.  But it also
has disadvantages.  For one thing, it reduces flexibility; if you change
the definition of the function, calls already inlined still use the old
definition until you recompile them.

Another disadvantage is that making a large function inline can increase
the size of compiled code both in files and in memory.  Since the
advantages of inline functions are greatest for small functions, you
generally should not make large functions inline.

Inline functions can be used and open coded later on in the same file,
following the definition, just like macros.

@item
The command @code{byte-compile-file} now offers to save any buffer
visiting the file you are compiling.

@item
The new command @code{compile-defun} reads, compiles and executes the
defun containing point.  If you use this on a defun that is actually a
function definition, the effect is to install a compiled version of
that function.

@item
Whenever you load a Lisp file or library, you now receive a warning if
the directory contains both a @samp{.el} file and a @samp{.elc} file,
and the @samp{.el} file is newer.  This typically indicates that someone
has updated the Lisp code but forgotten to recompile it, so the changes
do not take effect.  The warning is a reminder to recompile.

@item
The special form @code{eval-when-compile} marks the forms it contains to
be evaluated at compile time @emph{only}.  At top-level, this is
analogous to the Common Lisp idiom @code{(eval-when (compile)
@dots{})}.  Elsewhere, it is similar to the Common Lisp @samp{#.} reader
macro (but not when interpreting).

If you're thinking of using this feature, we recommend you consider whether
@code{provide} and @code{require} might do the job as well.

@item
The special form @code{eval-and-compile} is similar to
@code{eval-when-compile}, but the whole form is evaluated both at
compile time and at run time.

If you're thinking of using this feature, we recommend you consider
whether @code{provide} and @code{require} might do the job as well.

@item
Emacs Lisp has a new data type for byte-code functions.  This makes
them faster to call, and also saves space.  Internally, a byte-code
function object is much like a vector; however, the evaluator handles
this data type specially when it appears as a function to be called.

The printed representation for a byte-code function object is like that
for a vector, except that it starts with @samp{#} before the opening
@samp{[}.  A byte-code function object must have at least four elements;
there is no maximum number, but only the first six elements are actually
used.  They are:

@table @var
@item arglist
The list of argument symbols.

@item byte-code
The string containing the byte-code instructions.

@item constants
The vector of constants referenced by the byte code.

@item stacksize
The maximum stack size this function needs.

@item docstring
The documentation string (if any); otherwise, @code{nil}.

@item interactive
The interactive spec (if any).  This can be a string or a Lisp
expression.  It is @code{nil} for a function that isn't interactive.
@end table

The predicate @code{byte-code-function-p} tests whether a given object
is a byte-code function.

You can create a byte-code function object in a Lisp program
with the function @code{make-byte-code}.  Its arguments are the elements
to put in the byte-code function object.

You should not try to come up with the elements for a byte-code function
yourself, because if they are inconsistent, Emacs may crash when you
call the function.  Always leave it to the byte compiler to create these
objects; it, we hope, always makes the elements consistent.
@end itemize

@section Floating Point Numbers

You can now use floating point numbers in Emacs, if you define the macro
@code{LISP_FLOAT_TYPE} when you compile Emacs.

The printed representation for floating point numbers requires either a
decimal point surrounded by digits, or an exponent, or both.  For
example, @samp{1500.0}, @samp{15e2}, @samp{15.0e2} and @samp{1.5e3} are
four ways of writing a floating point number whose value is 1500.

The existing predicate @code{numberp} now returns @code{t} if the
argument is any kind of number---either integer or floating.  The new
predicates @code{integerp} and @code{floatp} check for specific types of
numbers.

You can do arithmetic on floating point numbers with the ordinary
arithmetic functions, @code{+}, @code{-}, @code{*} and @code{/}.  If you
call one of these functions with both integers and floating point
numbers among the arguments, the arithmetic is done in floating point.
The same applies to the numeric comparison functions such as @code{=}
and @code{<}.  The remainder function @code{%} does not accept floating
point arguments, and neither do the bitwise boolean operations such as
@code{logand} or the shift functions such as @code{ash}.

There is a new arithmetic function, @code{abs}, which returns the absolute
value of its argument.  It handles both integers and floating point
numbers.

To convert an integer to floating point, use the function @code{float}.
There are four functions to convert floating point numbers to integers;
they differ in how they round.  @code{truncate} rounds toward 0,
@code{floor} rounds down, @code{ceil} rounds up, and @code{round}
produces the nearest integer.

You can use @code{logb} to extract the binary exponent of a floating
point number.  More precisely, it is the logarithm base 2, rounded down
to an integer.

Emacs has several new mathematical functions that accept any kind of
number as argument, but always return floating point numbers.

@table @code
@item cos
@findex cos
@itemx sin
@findex sin
@itemx tan
@findex tan
Trigonometric functions.
@item acos
@findex acos
@itemx asin
@findex asin
@itemx atan
@findex atan
Inverse trigonometric functions.
@item exp
@findex exp
The exponential function (power of @var{e}).
@item log
@findex log
Logarithm base @var{e}.
@item expm1
@findex expm1
Power of @var{e}, minus 1.
@item log1p
@findex log1p
Add 1, then take the logarithm.
@item log10
@findex log10
Logarithm base 10
@item expt
@findex expt
Raise @var{x} to power @var{y}.
@item sqrt
@findex sqrt
The square root function.
@end table

The new function @code{string-to-number} now parses a string containing
either an integer or a floating point number, returning the number.

The @code{format} function now handles the specifications @samp{%e},
@samp{%f} and @samp{%g} for printing floating point numbers; likewise
@code{message}.

The new variable @code{float-output-format} controls how Lisp prints
floating point numbers.  Its value should be @code{nil} or a string.

If it is a string, it should contain a @samp{%}-spec like those accepted
by @code{printf} in C, but with some restrictions.  It must start with
the two characters @samp{%.}.  After that comes an integer which is the
precision specification, and then a letter which controls the format.

The letters allowed are @samp{e}, @samp{f} and @samp{g}.  Use @samp{e}
for exponential notation (@samp{@var{dig}.@var{digits}e@var{expt}}).
Use @samp{f} for decimal point notation
(@samp{@var{digits}.@var{digits}}).  Use @samp{g} to choose the shorter
of those two formats for the number at hand.

The precision in any of these cases is the number of digits following
the decimal point.  With @samp{f}, a precision of 0 means to omit the
decimal point.  0 is not allowed with @samp{f} or @samp{g}.

A value of @code{nil} means to use the format @samp{%.20g}.

No matter what the value of @code{float-output-format}, printing ensures
that the result fits the syntax rules for a floating point number.  If
it doesn't fit (for example, if it looks like an integer), it is
modified to fit.  By contrast, the @code{format} function formats
floating point numbers without requiring the output to fit the
syntax rules for floating point number.

@section New Features for Printing And Formatting Output

@itemize @bullet
@item
The @code{format} function has a new feature: @samp{%S}.  This print
spec prints any kind of Lisp object, even a string, using its Lisp
printed representation.

By contrast, @samp{%s} prints everything without quotation.

@item
@code{prin1-to-string} now takes an optional second argument which says
not to print the Lisp quotation characters.  (In other words, to use
@code{princ} instead of @code{prin1}.)

@item
The new variable @code{print-level} specifies the maximum depth of list
nesting to print before cutting off all deeper structure.  A value of
@code{nil} means no limit.
@end itemize

@section Changes in Basic Editing Functions

@itemize @bullet
@item
There are two new primitives for putting text in the kill ring:
@code{kill-new} and @code{kill-append}.

The function @code{kill-new} adds a string to the front of the kill ring.

Use @code{kill-append} to add a string to a previous kill.  The second
argument @var{before-p}, if non-@code{nil}, says to add the string at
the beginning; otherwise, it goes at the end.

Both of these functions apply @code{interprogram-cut-function} to the
entire string of killed text that ends up at the beginning of the kill
ring.

@item
The new function @code{current-kill} rotates the yanking pointer in the
kill ring by @var{n} places, and returns the text at that place in the
ring.  If the optional second argument @var{do-not-move} is
non-@code{nil}, it doesn't actually move the yanking point; it just
returns the @var{n}th kill forward.  If @var{n} is zero, indicating a
request for the latest kill, @code{current-kill} calls
@code{interprogram-paste-function} (documented below) before consulting
the kill ring.

All Emacs Lisp programs should either use @code{current-kill},
@code{kill-new}, and @code{kill-append} to manipulate the kill ring, or
be sure to call @code{interprogram-paste-function} and
@code{interprogram-cut-function} as appropriate.

@item
The variables @code{interprogram-paste-function} and
@code{interprogram-cut-function} exist so that you can provide functions
to transfer killed text to and from other programs.

@item
The @code{kill-region} function can now be used in read-only buffers.
It beeps, but adds the region to the kill ring without deleting it.

@item
The new function @code{compare-buffer-substrings} lets you compare two
substrings of the same buffer or two different buffers.  Its arguments
look like this:

@example
(compare-buffer-substrings @var{buf1} @var{beg1} @var{end1} @var{buf2} @var{beg2} @var{end2})
@end example

The first three arguments specify one substring, giving a buffer and two
positions within the buffer.  The last three arguments specify the other
substring in the same way.

The value is negative if the first substring is less, positive if the
first is greater, and zero if they are equal.  The absolute value of
the result is one plus the index of the first different characters.

@item
Overwrite mode treats tab and newline characters specially.  You can now
turn off this special treatment by setting @code{overwrite-binary-mode}
to @code{t}.

@item
Once the mark ``exists'' in a buffer, it normally never ceases to
exist.  However, it may become @dfn{inactive}.  The variable
@code{mark-active}, which is always local in all buffers, indicates
whether the mark is active: non-@code{nil} means yes.

A command can request deactivation of the mark upon return to the editor
command loop by setting @code{deactivate-mark} to a non-@code{nil}
value.  Transient Mark mode works by causing the buffer modification
primitives to set @code{deactivate-mark}.

The variables @code{activate-mark-hook} and @code{deactivate-mark-hook}
are normal hooks run, respectively, when the mark becomes active andwhen
it becomes inactive.  The hook @code{activate-mark-hook} is also run at
the end of a command if the mark is active and the region may have
changed.

@item
The function @code{move-to-column} now accepts a second optional
argument @var{force}, in addition to @var{column}; if the requested
column @var{column} is in the middle of a tab character and @var{force}
is non-@code{nil}, @code{move-to-column} replaces the tab with the
appropriate sequence of spaces so that it can place point exactly at
@var{column}.

@item
The search functions when successful now return the value of point
rather than just @code{t}.  This affects the functions
@code{search-forward}, @code{search-backward},
@code{word-search-forward}, @code{word-search-backward},
@code{re-search-forward}, and @code{re-search-backward}.

@item
When you do regular expression searching or matching, there is no longer
a limit to how many @samp{\(@dots{}\)} pairs you can get information
about with @code{match-beginning} and @code{match-end}.  Also, these
parenthetical groupings may now be nested to any degree.

@item
The new special form @code{save-match-data} preserves the regular
expression match status.  Usage: @code{(save-match-data
@var{body}@dots{})}.

@item
The function @code{translate-region} applies a translation table to the
characters in a part of the buffer.  Invoke it as
@code{(translate-region @var{start} @var{end} @var{table})}; @var{start}
and @var{end} bound the region to translate.

The translation table @var{table} is a string; @code{(aref @var{table}
@var{ochar})} gives the translated character corresponding to
@var{ochar}.  If the length of @var{table} is less than 256, any
characters with codes larger than the length of @var{table} are not
altered by the translation.

@code{translate-region} returns the number of characters which were
actually changed by the translation.  This does not count characters
which were mapped into themselves in the translation table.

@item
There are two new hook variables that let you notice all changes in all
buffers (or in a particular buffer, if you make them buffer-local):
@code{before-change-function} and @code{after-change-function}.

If @code{before-change-function} is non-@code{nil}, then it is called
before any buffer modification.  Its arguments are the beginning and end
of the region that is going to change, represented as integers.  The
buffer that's about to change is always the current buffer.

If @code{after-change-function} is non-@code{nil}, then it is called
after any buffer modification.  It takes three arguments: the beginning
and end of the region just changed, and the length of the text that
existed before the change.  (To get the current length, subtract the
rrgion beginning from the region end.)  All three arguments are
integers.  The buffer that's about to change is always the current
buffer.

Both of these variables are temporarily bound to @code{nil} during the
time that either of these hooks is running.  This means that if one of
these functions changes the buffer, that change won't run these
functions.  If you do want hooks to be run recursively, write your hook
functions to bind these variables back to their usual values.

@item
The hook @code{first-change-hook} is run using @code{run-hooks} whenever
a buffer is changed that was previously in the unmodified state.

@item
The second argument to @code{insert-abbrev-table-description} is
now optional.
@end itemize

@section Text Properties

  Each character in a buffer or a string can have a @dfn{text property
list}, much like the property list of a symbol.  The properties belong
to a particular character at a particular place, such as, the letter
@samp{T} at the beginning of this sentence.  Each property has a name,
which is usually a symbol, and an associated value, which can be any
Lisp object---just as for properties of symbols (@pxref{Property Lists}).

  You can use the property @code{face-code} to control the font and
color of text.  That is the only property name which currently has a
special meaning, but you can create properties of any name and examine
them later for your own purposes.

  Copying text between strings and buffers preserves the properties
along with the characters; this includes such diverse functions as
@code{substring}, @code{insert}, and @code{buffer-substring}.

  Since text properties are considered part of the buffer contents,
changing properties in a buffer ``modifies'' the buffer, and you can
also undo such changes.

  Strings with text properties have a special printed representation
which describes all the properties.  This representation is also the
read syntax for such a string.  It looks like this:

@example
#("@var{characters}" @var{property-data}...)
@end example

@noindent
where @var{property-data} is zero or more elements in groups of three as
follows:

@example
@var{beg} @var{end} @var{plist}
@end example

@noindent
The elements @var{beg} and @var{end} are integers, and together specify
a portion of the string; @var{plist} is the property list for that
portion.

@subsection Examining Text Properties

  The simplest way to examine text properties is to ask for the value of
a particular property of a particular character.  For that, use
@code{get-text-property}.  Use @code{text-properties-at} to get the
entire property list of a character.  @xref{Property Search}, for
functions to examine the properties of a number of characters at once.

@code{(get-text-property @var{pos} @var{prop} @var{object})} returns the
@var{prop} property of the character after @var{pos} in @var{object} (a
buffer or string).  The argument @var{object} is optional and defaults
to the current buffer.

@code{(text-properties-at @var{pos} @var{object})} returns the entire
property list of the character after @var{pos} in the string or buffer
@var{object} (which defaults to the current buffer).

@subsection Changing Text Properties

  There are three primitives for changing properties of a specified
range of text:

@table @code
@item add-text-properties
This function puts on specified properties, leaving other existing
properties unaltered.

@item put-text-property
This function puts on a single specified property, leaving others
unaltered.

@item remove-text-properties
This function removes specified properties, leaving other
properties unaltered.

@item set-text-properties
This function replaces the entire property list, leaving no vessage of
the properties that that text used to have.
@end table

All these functions take four arguments: @var{start}, @var{end},
@var{props}, and @var{object}.  The last argument is optional and
defaults to the current buffer.  The argument @var{props} has the form
of a property list.

@subsection Property Search Functions

In typical use of text properties, most of the time several or many
consecutive characters have the same value for a property.  Rather than
writing your programs to examine characters one by one, it is much
faster to process chunks of text that have the same property value.

The functions @code{next-property-change} and
@code{previous-property-change} scan forward or backward from position
@var{pos} in @var{object}, looking for a change in any property between
two characters scanned.  They returns the position between those two
characters, or @code{nil} if no change is found.

The functions @code{next-single-property-change} and
@code{previous-single-property-change} are similar except that you
specify a particular property and they look for changes in the value of
that property only.  The property is the second argument, and
@var{object} is third.

@subsection Special Properties

  If a character has a @code{category} property, we call it the
@dfn{category} of the character.  It should be a symbol.  The properties
of the symbol serve as defaults for the properties of the character.

  You can use the property @code{face-code} to control the font and
color of text.  That is the only property name which currently has a
special meaning, but you can create properties of any name and examine
them later for your own purposes.
about face codes.

  You can specify a different keymap for a portion of the text by means
of a @code{local-map} property.  The property's value, for the character
after point, replaces the buffer's local map.

  If a character has the property @code{read-only}, then modifying that
character is not allowed.  Any command that would do so gets an error.

  If a character has the property @code{modification-hooks}, then its
value should be a list of functions; modifying that character calls all
of those functions.  Each function receives two arguments: the beginning
and end of the part of the buffer being modified.  Note that if a
particular modification hook function appears on several characters
being modified by a single primitive, you can't predict how many times
the function will be called.

  Insertion of text does not, strictly speaking, change any existing
character, so there is a special rule for insertion.  It compares the
@code{read-only} properties of the two surrounding characters; if they
are @code{eq}, then the insertion is not allowed.  Assuming insertion is
allowed, it then gets the @code{modification-hooks} properties of those
characters and calls all the functions in each of them.  (If a function
appears on both characters, it may be called once or twice.)

  The special properties @code{point-entered} and @code{point-left}
record hook functions that report motion of point.  Each time point
moves, Emacs compares these two property values:

@itemize @bullet
@item
the @code{point-left} property of the character after the old location,
and
@item
the @code{point-entered} property of the character after the new
location.
@end itemize

@noindent
If these two values differ, each of them is called (if not @code{nil})
with two arguments: the old value of point, and the new one.

  The same comparison is made for the characters before the old and new
locations.  The result may be to execute two @code{point-left} functions
(which may be the same function) and/or two @code{point-entered}
functions (which may be the same function).  The @code{point-left}
functions are always called before the @code{point-entered} functions.

  A primitive function may examine characters at various positions
without moving point to those positions.  Only an actual change in the
value of point runs these hook functions.

@section New Features for Files

@itemize @bullet
@item
The new function @code{file-accessible-directory-p} tells you whether
you can open files in a particular directory.  Specify as an argument
either a directory name or a file name which names a directory file.
The function returns @code{t} if you can open existing files in that
directory.

@item
The new function @code{file-executable-p} returns @code{t} if its
argument is the name of a file you have permission to execute.

@item
The function @code{file-truename} returns the ``true name'' of a
specified file.  This is the name that you get by following symbolic
links until none remain.  The argument must be an absolute file name.

@item
New functions @code{make-directory} and @code{delete-directory} create and
delete directories.  They both take one argument, which is the name of
the directory as a file.

@item
The function @code{read-file-name} now takes an additional argument
which specifies an initial file name.  If you specify this argument,
@code{read-file-name} inserts it along with the directory name.  It puts
the cursor between the directory and the initial file name.

The user can then use the initial file name unchanged, modify it, or
simply kill it with @kbd{C-k}.

If the variable @code{insert-default-directory} is @code{nil}, then the
default directory is not inserted, and the new argument is ignored.

@item
The function @code{file-relative-name} does the inverse of
expansion---it tries to return a relative name which is equivalent to
@var{filename} when interpreted relative to @var{directory}.  (If such a
relative name would be longer than the absolute name, it returns the
absolute name instead.)

@item
The function @code{file-newest-backup} returns the name of the most
recent backup file for @var{filename}, or @code{nil} that file has no
backup files.

@item
The list returned by @code{file-attributes} now has 12 elements.  The
12th element is the file system number of the file system that the file
is in.  This element together with the file's inode number, which is the
11th element, give enough information to distinguish any two files on
the system---no two files can have the same values for both of these
numbers.

@item
The new function @code{set-visited-file-modtime} updates the current
buffer's recorded modification time from the visited file's time.

This is useful if the buffer was not read from the file normally, or
if the file itself has been changed for some known benign reason.

If you give the function an argument, that argument specifies the new
value for the recorded modification time.  The argument should be a list
of the form @code{(@var{high} . @var{low})} or @code{(@var{high}
@var{low})} containing two integers, each of which holds 16 bits of the
time.  (This is the same format that @code[file-attributes} uses to
return time values.)

The new function @code{visited-file-modtime} returns the recorded last
modification time, in that same format.

@item
The function @code{directory-files} now takes an optional fourth
argument which, if non-@code{nil}, inhibits sorting the file names.
Use this if you want the utmost possible speed and don't care what order
the files are processed in.

If the order of processing is at all visible to the user, then the user
will probably be happier if you do sort the names.

@item
The variable @code{directory-abbrev-alist} contains an alist of
abbreviations to use for file directories.  Each element has the form
@code{(@var{from} . @var{to})}, and says to replace @var{from} with
@var{to} when it appears in a directory name.  This replacement is done
when setting up the default directory of a newly visited file.  The
@var{from} string is actually a regular expression; it should always
start with @samp{^}.

You can set this variable in @file{site-init.el} to describe the
abbreviations appropriate for your site.

@item
The function @code{abbreviate-file-name} applies abbreviations from
@code{directory-abbrev-alist} to its argument, and substitutes @samp{~}
for the user's home directory.

Abbreviated directory names are useful for directories that are normally
accessed through symbolic links.  If you think of the link's name as
``the name'' of the directory, you can define it as an abbreviation for
the directory's official name; then ordinarily Emacs will call that
directory by the link name you normally use.

@item
@code{write-region} can write a given string instead of text from the
buffer.  Use the string as the first argument (in place of the
starting character position).

You can supply a second file name as the fifth argument (@var{visit}).
Use this to write the data to one file (the first argument,
@var{filename}) while nominally visiting a different file (the fifth
argument, @var{visit}).  The argument @var{visit} is used in the echo
area message and also for file locking; @var{visit} is stored in
@code{buffer-file-name}.

@item
The value of @code{write-file-hooks} does not change when you switch to
a new major mode.  The intention is that these hooks have to do with
where the file came from, and not with what it contains.

@item
There is a new hook variable for saving files:
@code{write-contents-hooks}.  It works just like @code{write-file-hooks}
except that switching to a new major mode clears it back to @code{nil}.
Major modes should use this hook variable rather than
@code{write-file-hooks}.

@item
The hook @code{after-save-hook} runs just after a buffer has been saved
in its visited file.

@item
The new function @code{set-default-file-modes} sets the file protection
for new files created with Emacs.  The argument must be an integer.  (It
would be better to permit symbolic arguments like the @code{chmod}
program, but that would take more work than this function merits.)

Use the new function @code{default-file-modes} to read the current
default file mode.

@item
Call the new function @code{unix-sync} to force all pending disk output
to happen as soon as possible.
@end itemize

@section Making Certain File Names ``Magic''

You can implement special handling for a class of file names.  You must
supply a regular expression to define the class of names (all those
which match the regular expression), plus a handler that implements all
the primitive Emacs file operations for file names that do match.

The value of @code{file-name-handler-alist} is a list of handlers,
together with regular expressions that decide when to apply each
handler.  Each element has the form @code{(@var{regexp}
. @var{handler})}.  If a file name matches @var{regexp}, then all work
on that file is done by calling @var{handler}.

All the Emacs primitives for file access and file name transformation
check the given file name against @code{file-name-handler-alist}, and
call @var{handler} to do the work if appropriate.  The first argument
given to @var{handler} is the name of the primitive; the remaining
arguments are the arguments that were passed to that primitive.  (The
first of these arguments is typically the file name itself.)  For
example, if you do this:

@example
(file-exists-p @var{filename})
@end example

@noindent
and @var{filename} has handler @var{handler}, then @var{handler} is
called like this:

@example
(funcall @var{handler} 'file-exists-p @var{filename})
@end example

Here are the primitives that you can handle in this way:

@quotation
@code{add-name-to-file}, @code{copy-file}, @code{delete-directory},
@code{delete-file}, @code{directory-file-name}, @code{directory-files},
@code{dired-compress-file}, @code{dired-uncache},
@code{expand-file-name}, @code{file-accessible-directory-p},
@code{file-attributes}, @code{file-directory-p},
@code{file-executable-p}, @code{file-exists-p}, @code{file-local-copy},
@code{file-modes}, @code{file-name-all-completions},
@code{file-name-as-directory}, @code{file-name-completion},
@code{file-name-directory}, @code{file-name-nondirectory},
@code{file-name-sans-versions}, @code{file-newer-than-file-p},
@code{file-readable-p}, @code{file-symlink-p}, @code{file-writable-p},
@code{insert-directory}, @code{insert-file-contents},
@code{make-directory}, @code{make-symbolic-link}, @code{rename-file},
@code{set-file-modes}, @code{verify-visited-file-modtime},
@code{write-region}.
@end quotation

The handler function must handle all of the above operations, and
possibly others to be added in the future.  Therefore, it should always
reinvoke the ordinary Lisp primitive when it receives an operation it
does not recognize.  Here's one way to do this:

@smallexample
(defun my-file-handler (primitive &rest args)
  ;; @r{First check for the specific operations}
  ;; @r{that we have special handling for.}
  (cond ((eq operation 'insert-file-contents) @dots{})
        ((eq operation 'write-region) @dots{})
        @dots{}
        ;; @r{Handle any operation we don't know about.}
        (t (let (file-name-handler-alist)
             (apply operation args)))))
@end smallexample

The function @code{file-local-copy} copies file @var{filename} to the
local site, if it isn't there already.  If @var{filename} specifies a
``magic'' file name which programs outside Emacs cannot directly read or
write, this copies the contents to an ordinary file and returns that
file's name.

If @var{filename} is an ordinary file name, not magic, then this function
does nothing and returns @code{nil}.

The function @code{unhandled-file-name-directory} is used to get a
non-magic directory name from an arbitrary file name.  It uses the
directory part of the specified file name if that is not magic.
Otherwise, it asks the file name's handler what to do.

@section Frames
@cindex frame

Emacs now supports multiple X windows via a new data type known as a
@dfn{frame}.

A frame is a rectangle on the screen that contains one or more Emacs
windows.  Subdividing a frame works just like subdividing the screen in
earlier versions of Emacs.

@cindex terminal frame
There are two kinds of frames: terminal frames and X window frames.
Emacs creates one terminal frame when it starts up with no X display; it
uses Termcap or Terminfo to display using characters.  There is no way
to create another terminal frame after startup.  If Emacs has an X
display, it does not make a terminal frame, and there is none.

@cindex X window frame
When you are using X windows, Emacs starts out with a single X window
frame.  You can create any number of X window frames using
@code{make-frame}.

Use the predicate @code{framep} to determine whether a given Lisp object
is a frame.

The function @code{redraw-frame} redisplays the entire contents of a
given frame.

@subsection Creating and Deleting Frames

Use @code{make-frame} to create a new frame (supported under X Windows
only).  This is the only primitive for creating frames.

@code{make-frame} takes just one argument, which is an alist
specifying frame parameters.  Any parameters not mentioned in the
argument alist default based on the value of @code{default-frame-alist};
parameters not specified there default from the standard X defaults file
and X resources.

When you invoke Emacs, if you specify arguments for window appearance
and so forth, these go into @code{default-frame-alist} and that is how
they have their effect.

You can specify the parameters for the initial startup X window frame by
setting @code{initial-frame-alist} in your @file{.emacs} file.  If these
parameters specify a separate minibuffer-only frame, and you have not
created one, Emacs creates one for you, using the parameter values
specified in @code{minibuffer-frame-alist}.

You can specify the size and position of a frame using the frame
parameters @code{left}, @code{top}, @code{height} and @code{width}.  You
must specify either both size parameters or neither.  You must specify
either both position parameters or neither.  The geometry parameters
that you don't specify are chosen by the window manager in its usual
fashion.

The function @code{x-parse-geometry} converts a standard X windows
geometry string to an alist which you can use as part of the argument to
@code{make-frame}.

Use the function @code{delete-frame} to eliminate a frame.  Frames are
like buffers where deletion is concerned; a frame actually continues to
exist as a Lisp object until it is deleted @emph{and} there are no
references to it, but once it is deleted, it has no further effect on
the screen.

The function @code{frame-live-p} returns non-@code{nil} if the argument
(a frame) has not been deleted.

@subsection Finding All Frames

The function @code{frame-list} returns a list of all the frames that have
not been deleted.  It is analogous to @code{buffer-list}.  The list that
you get is newly created, so modifying the list doesn't have any effect
on the internals of Emacs.  The function @code{visible-frame-list} returns
the list of just the frames that are visible.

@code{next-frame} lets you cycle conveniently through all the frames from an
arbitrary starting point.  Its first argument is a frame.  Its second
argument @var{minibuf} says what to do about minibuffers:

@table @asis
@item @code{nil}
Exclude minibuffer-only frames.
@item a window
Consider only the frames using that particular window as their
minibuffer.
@item anything else
Consider all frames.
@end table

@subsection Frames and Windows

All the non-minibuffer windows in a frame are arranged in a tree of
subdivisions; the root of this tree is available via the function
@code{frame-root-window}.  Each window is part of one and only one
frame; you can get the frame with @code{window-frame}.

At any time, exactly one window on any frame is @dfn{selected within the
frame}.  You can get the frame's current selected window with
@code{frame-selected-window}.  The significance of this designation is
that selecting the frame selects for Emacs as a whole the window
currently selected within that frame.

Conversely, selecting a window for Emacs with @code{select-window} also
makes that window selected within its frame.

@subsection Frame Visibility

A frame may be @dfn{visible}, @dfn{invisible}, or @dfn{iconified}.  If
it is invisible, it doesn't show in the screen, not even as an icon.
You can set the visibility status of a frame with
@code{make-frame-visible}, @code{make-frame-invisible}, and
@code{iconify-frame}.  You can examine the visibility status with
@code{frame-visible-p}---it returns @code{t} for a visible frame,
@code{nil} for an invisible frame, and @code{icon} for an iconified
frame.

@subsection Selected Frame

At any time, one frame in Emacs is the @dfn{selected frame}.  The selected
window always resides on the selected frame.

@defun selected-frame
This function returns the selected frame.
@end defun

The X server normally directs keyboard input to the X window that the
mouse is in.  Some window managers use mouse clicks or keyboard events
to @dfn{shift the focus} to various X windows, overriding the normal
behavior of the server.

Lisp programs can switch frames ``temporarily'' by calling the function
@code{select-frame}.  This does not override the window manager; rather,
it escapes from the window manager's control until that control is
somehow reasserted.  The function takes one argument, a frame, and
selects that frame.  The selection lasts until the next time the user
does something to select a different frame, or until the next time this
function is called.

Emacs cooperates with the X server and the window managers by arranging
to select frames according to what the server and window manager ask
for.  It does so by generating a special kind of input event, called a
@dfn{focus} event.  The command loop handles a focus event by calling
@code{internal-select-frame}.  @xref{Focus Events}.

@subsection Frame Size and Position

The new functions @code{frame-height} and @code{frame-width} return the
height and width of a specified frame (or of the selected frame),
measured in characters.

The new functions @code{frame-pixel-height} and @code{frame-pixel-width}
return the height and width of a specified frame (or of the selected
frame), measured in pixels.

The new functions @code{frame-char-height} and @code{frame-char-width}
return the height and width of a character in a specified frame (or in
the selected frame), measured in pixels.

@code{set-frame-size} sets the size of a frame, measured in characters;
its arguments are @var{frame}, @var{cols} and @var{rows}.  To set the
size with values measured in pixels, you can use
@code{modify-frame-parameters}.

The function @code{set-frame-position} sets the position of the top left
corner of a frame.  Its arguments are @var{frame}, @var{left} and
@var{top}.

@ignore
New functions @code{set-frame-height} and @code{set-frame-width} set the
size of a specified frame.  The frame is the first argument; the size is
the second.
@end ignore

@subsection Frame Parameters

A frame has many parameters that affect how it displays.  Use the
function @code{frame-parameters} to get an alist of all the parameters
of a given frame.  To alter parameters, use
@code{modify-frame-parameters}, which takes two arguments: the frame to
modify, and an alist of parameters to change and their new values.  Each
element of @var{alist} has the form @code{(@var{parm} . @var{value})},
where @var{parm} is a symbol.  Parameters that aren't meaningful are
ignored.  If you don't mention a parameter in @var{alist}, its value
doesn't change.

Just what parameters a frame has depends on what display mechanism it
uses.  Here is a table of the parameters of an X
window frame:

@table @code
@item name
The name of the frame.

@item left
The screen position of the left edge.

@item top
The screen position of the top edge.

@item height
The height of the frame contents, in pixels.

@item width
The width of the frame contents, in pixels.

@item window-id
The number of the X window for the frame.

@item minibuffer
Whether this frame has its own minibuffer.
@code{t} means yes, @code{none} means no,
@code{only} means this frame is just a minibuffer,
a minibuffer window (in some other frame)
means the new frame uses that minibuffer.

@item font
The name of the font for the text.

@item foreground-color
The color to use for the inside of a character.
Use strings to designate colors;
X windows defines the meaningful color names.

@item background-color
The color to use for the background of text.

@item mouse-color
The color for the mouse cursor.

@item cursor-color
The color for the cursor that shows point.

@item border-color
The color for the border of the frame.

@item cursor-type
The way to display the cursor.  There are two legitimate values:
@code{bar} and @code{box}.  The value @code{bar} specifies a vertical
bar between characters as the cursor.  The value @code{box} specifies an
ordinary black box overlaying the character after point; that is the
default.

@item icon-type
Non-@code{nil} for a bitmap icon, @code{nil} for a text icon.

@item border-width
The width in pixels of the window border.

@item internal-border-width
The distance in pixels between text and border.

@item auto-raise
Non-@code{nil} means selecting the frame raises it.

@item auto-lower
Non-@code{nil} means deselecting the frame lowers it.

@item vertical-scrollbar
Non-@code{nil} gives the frame a scroll bar
for vertical scrolling.

@item horizontal-scrollbar
Non-@code{nil} gives the frame a scroll bar
for horizontal scrolling.
@end table

@subsection Minibufferless Frames

Normally, each frame has its own minibuffer window at the bottom, which
is used whenever that frame is selected.  However, you can also create
frames with no minibuffers.  These frames must use the minibuffer window
of some other frame.

The variable @code{default-minibuffer-frame} specifies where to find a
minibuffer for frames created without minibuffers of their own.  Its
value should be a frame which does have a minibuffer.

You can also specify a minibuffer window explicitly when you create a
frame; then @code{default-minibuffer-frame} is not used.

@section X Windows Features

@itemize @bullet
@item
The new functions @code{mouse-position} and @code{set-mouse-position} give
access to the current position of the mouse.

@code{mouse-position} returns a description of the position of the mouse.
The value looks like @code{(@var{frame} @var{x} . @var{y})}, where @var{x}
and @var{y} are measured in pixels relative to the top left corner of
the inside of @var{frame}.

@code{set-mouse-position} takes three arguments, @var{frame}, @var{x}
and @var{y}, and warps the mouse cursor to that location on the screen.

@item
@code{track-mouse} is a new special form for tracking mouse motion.
Use it in definitions of mouse clicks that want pay to attention to
the motion of the mouse, not just where the buttons are pressed and
released.  Here is how to use it:

@example
(track-mouse @var{body}@dots{})
@end example

While @var{body} executes, mouse motion generates input events just as mouse
clicks do.  @var{body} can read them with @code{read-event} or
@code{read-key-sequence}.

@code{track-mouse} returns the value of the last form in @var{body}.

The format of these events is described under ``New features for key
bindings and input.''
@c ???

@item
@code{x-set-selection} sets a ``selection'' in the X Windows server.
It takes two arguments: a selection type @var{type}, and the value to
assign to it, @var{data}.  If @var{data} is @code{nil}, it means to
clear out the selection.  Otherwise, @var{data} may be a string, a
symbol, an integer (or a cons of two integers or list of two integers),
or a cons of two markers pointing to the same buffer.  In the last case,
the selection is considered to be the text between the markers.  The
data may also be a vector of valid non-vector selection values.

Each possible @var{type} has its own selection value, which changes
independently.  The usual values of @var{type} are @code{PRIMARY} and
@code{SECONDARY}; these are symbols with upper-case names, in accord
with X Windows conventions.  The default is @code{PRIMARY}.

To get the value of the selection, call @code{x-get-selection}.  This
function accesses selections set up by Emacs and those set up by other X
clients.  It takes two optional arguments, @var{type} and
@var{data-type}.  The default for @var{type} is @code{PRIMARY}.

The @var{data-type} argument specifies the form of data conversion to
use; meaningful values include @code{TEXT}, @code{STRING},
@code{TARGETS}, @code{LENGTH}, @code{DELETE}, @code{FILE_NAME},
@code{CHARACTER_POSITION}, @code{LINE_NUMBER}, @code{COLUMN_NUMBER},
@code{OWNER_OS}, @code{HOST_NAME}, @code{USER}, @code{CLASS},
@code{NAME}, @code{ATOM}, and @code{INTEGER}.  (These are symbols with
upper-case names in accord with X Windows conventions.)
The default for @var{data-type} is @code{STRING}.

@item
X Windows has a set of numbered @dfn{cut buffers} which can store text
or other data being moved between applications.  Use
@code{x-get-cut-buffer} to get the contents of a cut buffer; specify the
cut buffer number as argument.  Use @code{x-set-cut-buffer} with
argument @var{string} to store a new string into the first cut buffer
(moving the other values down through the series of cut buffers,
kill-ring-style).

Cut buffers are considered obsolete in X Windows, but Emacs supports
them for the sake of X clients that still use them.

@item
You can close the connection with the X Windows server with
the function @code{x-close-current-connection}.  This takes no arguments.

Then you can connect to a different X Windows server with
@code{x-open-connection}.  The first argument, @var{display}, is the
name of the display to connect to.

The optional second argument @var{xrm-string} is a string of resource
names and values, in the same format used in the @file{.Xresources}
file.  The values you specify override the resource values recorded in
the X Windows server itself.  Here's an example of what this string
might look like:

@example
"*BorderWidth: 3\n*InternalBorder: 2\n"
@end example

@item
A series of new functions give you information about the X server and
the screen you are using.

@table @code
@item x-display-screens
The number of screens associated with the current display.

@item x-server-version
The version numbers of the X server in use.

@item x-server-vendor
The vendor supporting the X server in use.

@item x-display-pixel-height
The height of this X screen in pixels.

@item x-display-mm-height
The height of this X screen in millimeters.

@item x-display-pixel-width
The width of this X screen in pixels.

@item x-display-mm-width
The width of this X screen in millimeters.

@item x-display-backing-store
The backing store capability of this screen.  Values can be the symbols
@code{always}, @code{when-mapped}, or @code{not-useful}.

@item x-display-save-under
Non-@code{nil} if this X screen supports the SaveUnder feature.

@item x-display-planes
The number of planes this display supports.

@item x-display-visual-class
The visual class for this X screen.  The value is one of the symbols
@code{static-gray}, @code{gray-scale}, @code{static-color},
@code{pseudo-color}, @code{true-color}, and @code{direct-color}.

@item x-display-color-p
@code{t} if the X screen in use is a color screen.

@item x-display-color-cells
The number of color cells this X screen supports.
@end table

There is also a variable @code{x-no-window-manager}, whose value is
@code{t} if no X window manager is in use.

@item
The function @code{x-synchronize} enables or disables an X Windows
debugging mode: synchronous communication.  It takes one argument,
non-@code{nil} to enable the mode and @code{nil} to disable.

In synchronous mode, Emacs waits for a response to each X protocol
command before doing anything else.  This means that errors are reported
right away, and you can directly find the erroneous command.
Synchronous mode is not the default because it is much slower.

@item
The function @code{x-get-resource} retrieves a resource value from the X
Windows defaults database.  Its three arguments are @var{attribute},
@var{name} and @var{class}.  It searches using a key of the form
@samp{@var{instance}.@var{attribute}}, with class @samp{Emacs}, where
@var{instance} is the name under which Emacs was invoked.

The optional arguments @var{component} and @var{subclass} add to the key
and the class, respectively.  You must specify both of them or neither.
If you specify them, the key is
@samp{@var{instance}.@var{component}.@var{attribute}}, and the class is
@samp{Emacs.@var{subclass}}.

@item
@code{x-color-display-p} returns @code{t} if you are using an X Window
server with a color display, and @code{nil} otherwise.

@c ??? Name being changed from x-defined-color.
@code{x-color-defined-p} takes as argument a string describing a color; it
returns @code{t} if the display supports that color.  (If the color is
@code{"black"} or @code{"white"} then even black-and-white displays
support it.)

@item
@code{x-popup-menu} has been generalized.  It now accepts a keymap as
the @var{menu} argument.  Then the menu items are the prompt strings of
individual key bindings, and the item values are the keys which have
those bindings.

You can also supply a list of keymaps as the first argument; then each
keymap makes one menu pane (but keymaps that don't provide any menu
items don't appear in the menu at all).

@code{x-popup-menu} also accepts a mouse button event as the
@var{position} argument.  Then it displays the menu at the location at
which the event took place.  This is convenient for mouse-invoked
commands that pop up menus.

@ignore
@item
x-pointer-shape, x-nontext-pointer-shape, x-mode-pointer-shape.
@end ignore

@item
You can use the function @code{x-rebind-key} to change the sequence
of characters generated by one of the keyboard keys.  This works
only with X Windows.

The first two arguments, @var{keycode} and @var{shift-mask}, should be
numbers representing the keyboard code and shift mask respectively.
They specify what key to change.

The third argument, @var{newstring}, is the new definition of the key.
It is a sequence of characters that the key should produce as input.

The shift mask value is a combination of bits according to this table:

@table @asis
@item 8
Control
@item 4
Meta
@item 2
Shift
@item 1
Shift Lock
@end table

If you specify @code{nil} for @var{shift-mask}, then the key specified
by @var{keycode} is redefined for all possible shift combinations.

For the possible values of @var{keycode} and their meanings, see the
file @file{/usr/lib/Xkeymap.txt}.  Keep in mind that the codes in that
file are in octal!

@ignore @c Presumably this is already fixed
NOTE: due to an X bug, this function will not take effect unless the
user has a @file{~/.Xkeymap} file.  (See the documentation for the
@code{keycomp} program.)  This problem will be fixed in X version 11.
@end ignore

The related function @code{x-rebind-keys} redefines a single keyboard
key, specifying the behavior for each of the 16 shift masks
independently.  The first argument is @var{keycode}, as in
@code{x-rebind-key}.  The second argument @var{strings} is a list of 16
elements, one for each possible shift mask value; each element says how
to redefine the key @var{keycode} with the corresponding shift mask
value.  If an element is a string, it is the new definition.  If an
element is @code{nil}, the definition does not change for that shift
mask.

@item
The function @code{x-geometry} parses a string specifying window size
and position in the usual fashion for X windows.  It returns an alist
describing which parameters were specified, and the values that were
given for them.

The elements of the alist look like @code{(@var{parameter} .
@var{value})}.  The possible @var{parameter} values are @code{left},
@code{top}, @code{width}, and @code{height}.
@end itemize

@section New Window Features

@itemize @bullet
@item
The new function @code{window-at} tells you which window contains a
given horizontal and vertical position on a specified frame.  Call it
with three arguments, like this:

@example
(window-at @var{x} @var{column} @var{frame})
@end example

The function returns the window which contains that cursor position in
the frame @var{frame}.  If you omit @var{frame}, the selected frame is
used.

@item
The function @code{coordinates-in-window-p} takes two arguments and
checks whether a particular frame position falls within a particular
window.

@example
(coordinates-in-window-p @var{coordinates} @var{window})
@end example

The argument @var{coordinates} is a cons cell of this form:

@example
(@var{x} . @var{y})
@end example

@noindent
The two coordinates are measured in characters, and count from the top
left corner of the screen or frame.

The value of the function tells you what part of the window the position
is in.  The possible values are:

@table @code
@item (@var{relx} . @var{rely})
The coordinates are inside @var{window}.  The numbers @var{relx} and
@var{rely} are equivalent window-relative coordinates, counting from 0
at the top left corner of the window.

@item mode-line
The coordinates are in the mode line of @var{window}.

@item vertical-split
The coordinates are in the vertical line between @var{window} and its
neighbor to the right.

@item nil
The coordinates are not in any sense within @var{window}.
@end table

You need not specify a frame when you call
@code{coordinates-in-window-p}, because it assumes you mean the frame
which window @var{window} is on.

@item
The function @code{minibuffer-window} now accepts a frame as argument
and returns the minibuffer window used for that frame.  If you don't
specify a frame, the currently selected frame is used.  The minibuffer
window may be on the frame in question, but if that frame has no
minibuffer of its own, it uses the minibuffer window of some other
frame, and @code{minibuffer-window} returns that window.

@item
Use @code{window-live-p} to test whether a window is still alive (that
is, not deleted).

@item
Use @code{window-minibuffer-p} to determine whether a given window is a
minibuffer or not.  It no longer works to do this by comparing the
window with the result of @code{(minibuffer-window)}, because there can
be more than one minibuffer window at a time (if you have multiple
frames).

@item
If you set the variable @code{pop-up-frames} non-@code{nil}, then the
functions to show something ``in another window'' actually create a new
frame for the new window.  Thus, you will tend to have a frame for each
window, and you can easily have a frame for each buffer.

The value of the variable @code{pop-up-frame-function} controls how new
frames are made.  The value should be a function which takes no
arguments and returns a frame.  The default value is a function which
creates a frame using parameters from @code{pop-up-frame-alist}.

@item
@code{display-buffer} is the basic primitive for finding a way to show a
buffer on the screen.  You can customize its behavior by storing a
function in the variable @code{display-buffer-function}.  If this
variable is non-@code{nil}, then @code{display-buffer} calls it to do
the work.  Your function should accept two arguments, as follows:

@table @var
@item buffer
The buffer to be displayed.

@item flag
A flag which, if non-@code{nil}, means you should find another window to
display @var{buffer} in, even if it is already visible in the selected
window.
@end table

The function you supply will be used by commands such as
@code{switch-to-buffer-other-window} and @code{find-file-other-window}
as well as for your own calls to @code{display-buffer}.

@item
@code{delete-window} now gives all of the deleted window's screen space
to a single neighboring window.  Likewise, @code{enlarge-window} takes
space from only one neighboring window until that window disappears;
only then does it take from another window.

@item
@code{next-window} and @code{previous-window} accept another argument,
@var{all-frames}.

These functions now take three optional arguments: @var{window},
@var{minibuf} and @var{all-frames}.  @var{window} is the window to start
from (@code{nil} means use the selected window).  @var{minibuf} says
whether to include the minibuffer in the windows to cycle through:
@code{t} means yes, @code{nil} means yes if it is active, and anything
else means no.

Normally, these functions cycle through all the windows in the
selected frame, plus the minibuffer used by the selected frame even if
it lies in some other frame.

If @var{all-frames} is @code{t}, then these functions cycle through
all the windows in all the frames that currently exist.  If
@var{all-frames} is neither @code{t} nor @code{nil}, then they limit
themselves strictly to the windows in the selected frame, excluding the
minibuffer in use if it lies in some other frame.

@item
The functions @code{get-lru-window} and @code{get-largest-window} now
take an optional argument @var{all-frames}.  If it is non-@code{nil},
the functions consider all windows on all frames.  Otherwise, they
consider just the windows on the selected frame.

Likewise, @code{get-buffer-window} takes an optional second argument
@var{all-frames}.

@item
The variable @code{other-window-scroll-buffer} specifies which buffer
@code{scroll-other-window} should scroll.

@item
You can now mark a window as ``dedicated'' to its buffer.
Then Emacs will not try to use that window for any other buffer
unless you explicitly request it.

Use the new function @code{set-window-dedicated-p} to set the dedication
flag of a window @var{window} to the value @var{flag}.  If @var{flag} is
@code{t}, this makes the window dedicated.  If @var{flag} is
@code{nil}, this makes the window non-dedicated.

Use @code{window-dedicated-p} to examine the dedication flag of a
specified window.

@item
The new function @code{walk-windows} cycles through all visible
windows, calling @code{proc} once for each window with the window as
its sole argument.

The optional second argument @var{minibuf} says whether to include minibuffer
windows.  A value of @code{t} means count the minibuffer window even if
not active.  A value of @code{nil} means count it only if active.  Any
other value means not to count the minibuffer even if it is active.

If the optional third argument @var{all-frames} is @code{t}, that means
include all windows in all frames.  If @var{all-frames} is @code{nil},
it means to cycle within the selected frame, but include the minibuffer
window (if @var{minibuf} says so) that that frame uses, even if it is on
another frame.  If @var{all-frames} is neither @code{nil} nor @code{t},
@code{walk-windows} sticks strictly to the selected frame.

@item
The function @code{window-end} is a counterpart to @code{window-start}:
it returns the buffer position of the end of the display in a given
window (or the selected window).

@item
The function @code{window-configuration-p} returns non-@code{nil} when
given an object that is a window configuration (such as is returned by
@code{current-window-configuration}).
@end itemize

@section Display Features

@itemize @bullet
@item
@samp{%l} as a mode line item displays the current line number.

If the buffer is longer than @code{line-number-display-limit}
characters, or if lines are too long in the viscinity of the current
displayed text, then line number display is inhibited to save time.

The default contents of the mode line include the line number if
@code{line-number-mode} is non-@code{nil}.

@item
@code{baud-rate} is now a variable rather than a function.  This is so
you can set it to reflect the effective speed of your terminal, when the
system doesn't accurately know the speed.

@item
You can now remove any echo area message and make the minibuffer
visible.  To do this, call @code{message} with @code{nil} as the only
argument.  This clears any existing message, and lets the current
minibuffer contents show through.  Previously, there was no reliable way
to make sure that the minibuffer contents were visible.

@item
The variable @code{temp-buffer-show-hook} has been renamed
@code{temp-buffer-show-function}, because its value is a single function
(of one argument), not a normal hook.

@item
The new function @code{force-mode-line-update} causes redisplay
of the current buffer's mode line.
@end itemize

@section Display Tables

@cindex display table
You can use the @dfn{display table} feature to control how all 256
possible character codes display on the screen.  This is useful for
displaying European languages that have letters not in the ASCII
character set.

The display table maps each character code into a sequence of
@dfn{glyphs}, each glyph being an image that takes up one character
position on the screen.  You can also define how to display each glyph
on your terminal, using the @dfn{glyph table}.

@subsection Display Tables

Use @code{make-display-table} to create a display table.  The table
initially has @code{nil} in all elements.

A display table is actually an array of 261 elements.  The first 256
elements of a display table control how to display each possible text
character.  The value should be @code{nil} or a vector (which is a
sequence of glyphs; see below).  @code{nil} as an element means to
display that character following the usual display conventions.

The remaining five elements of a display table serve special purposes
(@code{nil} means use the default stated below):

@table @asis
@item 256
The glyph for the end of a truncated screen line (the default for this
is @samp{\}).
@item 257
The glyph for the end of a continued line (the default is @samp{$}).
@item 258
The glyph for the indicating an octal character code (the default is
@samp{\}).
@item 259
The glyph for indicating a control characters (the default is @samp{^}).
@item 260
The vector of glyphs for indicating the presence of invisible lines (the
default is @samp{...}).
@end table

Each buffer typically has its own display table.  The display table for
the current buffer is stored in @code{buffer-display-table}.  (This
variable automatically becomes local if you set it.)  If this variable
is @code{nil}, the value of @code{standard-display-table} is used in
that buffer.

Each window can have its own display table, which overrides the display
table of the buffer it is showing.

If neither the selected window nor the current buffer has a display
table, and if @code{standard-display-table} is @code{nil}, then Emacs
uses the usual display conventions:

@itemize @bullet
@item
Character codes 32 through 127 map to glyph codes 32 through 127.
@item
Codes 0 through 31 map to sequences of two glyphs, where the first glyph
is the ASCII code for @samp{^}.
@item
Character codes 128 through 255 map to sequences of four glyphs, where
the first glyph is the ASCII code for @samp{\}, and the others represent
digits.
@end itemize

The usual display conventions are also used for any character whose
entry in the active display table is @code{nil}.  This means that when
you set up a display table, you need not specify explicitly what to do
with each character, only the characters for which you want unusual
behavior.

@subsection Glyphs

@cindex glyph
A glyph stands for an image that takes up a single character position on
the screen.  A glyph is represented in Lisp as an integer.

@cindex glyph table
The meaning of each integer, as a glyph, is defined by the glyph table,
which is the value of the variable @code{glyph-table}.  It should be a
vector; the @var{g}th element defines glyph code @var{g}.  The possible
definitions of a glyph code are:

@table @var
@item integer
Define this glyph code as an alias for code @var{integer}.
This is used with X windows to specify a face code.

@item string
Send the characters in @var{string} to the terminal to output
this glyph.  This alternative is not available with X Windows.

@item @code{nil}
This glyph is simple.  On an ordinary terminal, the glyph code mod 256
is the character to output.  With X, the glyph code mod 256 is character
to output, and the glyph code divided by 256 specifies the @dfn{face
code} to use while outputting it.
@end table

Any glyph code beyond the length of the glyph table is automatically simple.

A face code for X windows is the combination of a font and a color.
Emacs uses integers to identify face codes.  You can define a new face
code with @code{(x-set-face @var{face-code} @var{font} @var{foreground}
@var{background})}.  @var{face-code} is an integer from 0 to 255; it
specifies which face to define.  The other three arguments are strings:
@var{font} is the name of the font to use, and @var{foreground} and
@var{background} specify the colors to use.

If @code{glyph-table} is @code{nil}, then all possible glyph codes are
simple.

@subsection ISO Latin 1

If you have a terminal that can handle the entire ISO Latin 1 character
set, you can arrange to use that character set as follows:

@example
(require 'disp-table)
(standard-display-8bit 0 255)
@end example

If you are editing buffers written in the ISO Latin 1 character set and
your terminal doesn't handle anything but ASCII, you can load the file
@code{iso-ascii} to set up a display table which makes the other ISO
characters display as sequences of ASCII characters.  For example, the
character ``o with umlaut'' displays as @samp{@{"o@}}.

Some European countries have terminals that don't support ISO Latin 1
but do support the special characters for that country's language.  You
can define a display table to work one language using such terminals.
For an example, see @file{lisp/iso-swed.el}, which handles certain
Swedish terminals.

You can load the appropriate display table for your terminal
automatically by writing a terminal-specific Lisp file for the terminal
type.

@section New Input Event Formats

Mouse clicks, mouse movements and function keys no longer appear in the
input stream as characters; instead, other kinds of Lisp objects
represent them as input.

@itemize @bullet
@item
An ordinary input character event consists of a @dfn{basic code} between
0 and 255, plus any or all of these @dfn{modifier bits}:

@table @asis
@item meta
The 2**23 bit in the character code indicates a character
typed with the meta key held down.

@item control
The 2**22 bit in the character code indicates a non-@sc{ASCII}
control character.

@sc{ASCII} control characters such as @kbd{C-a} have special basic
codes of their own, so Emacs needs no special bit to indicate them.
Thus, the code for @kbd{C-a} is just 1.

But if you type a control combination not in @sc{ASCII}, such as
@kbd{%} with the control key, the numeric value you get is the code
for @kbd{%} plus 2**22 (assuming the terminal supports non-@sc{ASCII}
control characters).

@item shift
The 2**21 bit in the character code indicates an @sc{ASCII} control
character typed with the shift key held down.

For letters, the basic code indicates upper versus lower case; for
digits and punctuation, the shift key selects an entirely different
character with a different basic code.  In order to keep within
the @sc{ASCII} character set whenever possible, Emacs avoids using
the 2**21 bit for those characters.

However, @sc{ASCII} provides no way to distinguish @kbd{C-A} from
@kbd{C-A}, so Emacs uses the 2**21 bit in @kbd{C-A} and not in
@kbd{C-a}.

@item hyper
The 2**20 bit in the character code indicates a character
typed with the hyper key held down.

@item super
The 2**19 bit in the character code indicates a character
typed with the super key held down.

@item alt
The 2**18 bit in the character code indicates a character typed with
the alt key held down.  (On some terminals, the key labeled @key{ALT}
is actually the meta key.)
@end table

In the future, Emacs may support a larger range of basic codes.  We may
also move the modifier bits to larger bit numbers.  Therefore, you
should avoid mentioning specific bit numbers in your program.  Instead,
the way to test the modifier bits of a character is with the function
@code{event-modifiers} (see below).

@item
Function keys are represented as symbols.  The symbol's name is
the function key's label.  For example, pressing a key labeled @key{F1}
places the symbol @code{f1} in the input stream.

There are a few exceptions to the symbol naming convention:

@table @asis
@item @code{kp-add}, @code{kp-decimal}, @code{kp-divide}, @dots{}
Keypad keys (to the right of the regular keyboard).
@item @code{kp-0}, @code{kp-1}, @dots{}
Keypad keys with digits.
@item @code{kp-f1}, @code{kp-f2}, @code{kp-f3}, @code{kp-f4}
Keypad PF keys.
@item @code{left}, @code{up}, @code{right}, @code{down}
Cursor arrow keys
@end table

You can use the modifier keys @key{CTRL}, @key{META}, @key{HYPER},
@key{SUPER}, @key{ALT} and @key{SHIFT} with function keys.  The way
to represent them is with prefixes in the symbol name:

@table @samp
@item A-
The alt modifier.
@item C-
The control modifier.
@item H-
The hyper modifier.
@item M-
The meta modifier.
@item s-
The super modifier.
@item S-
The shift modifier.
@end table

Thus, the symbol for the key @key{F3} with @key{META} held down is
kbd{M-@key{F3}}.  When you use more than one prefix, we recommend you
write them in alphabetical order (though the order does not matter in
arguments to the key-binding lookup and modification functions).

@item
Mouse events are represented as lists.

If you press a mouse button and release it at the same location, this
generates a ``click'' event.  Mouse click events have this form:

@example
(@var{button-symbol}
 (@var{window} (@var{column} . @var{row})
  @var{buffer-pos} @var{timestamp}))
@end example

Here is what the elements normally mean:

@table @var
@item button-symbol
indicates which mouse button was used.  It is one of the symbols
@code{mouse-1}, @code{mouse-2}, @dots{}, where the buttons are numbered
numbered left to right.

You can also use prefixes @samp{A-}, @samp{C-}, @samp{H-}, @samp{M-},
@samp{S-} and @samp{s-} for modifiers alt, control, hyper, meta, shift
and super, just as you would with function keys.

@item window
is the window in which the click occurred.

@item column
@itemx row
are the column and row of the click, relative to the top left corner of
@var{window}, which is @code{(0 . 0)}.

@item buffer-pos
is the buffer position of the character clicked on.

@item timestamp
is the time at which the event occurred, in milliseconds.  (Since this
value wraps around the entire range of Emacs Lisp integers in about five
hours, it is useful only for relating the times of nearby events.)
@end table

The meanings of @var{buffer-pos}, @var{row} and @var{column} are
somewhat different when the event location is in a special part of the
screen, such as the mode line or a scroll bar.

If the position is in the window's scroll bar, then @var{buffer-pos} is
the symbol @code{vertical-scrollbar} or @code{horizontal-scrollbar}, and
the pair @code{(@var{column} . @var{row})} is instead a pair
@code{(@var{portion} . @var{whole})}, where @var{portion} is the
distance of the click from the top or left end of the scroll bar, and
@var{whole} is the length of the entire scroll bar.

If the position is on a mode line or the vertical line separating
@var{window} from its neighbor to the right, then @var{buffer-pos} is
the symbol @code{mode-line} or @code{vertical-line}.  In this case
@var{row} and @var{column} do not have meaningful data.

@item
Releasing a mouse button above a different character position
generates a ``drag'' event, which looks like this:

@example
(@var{button-symbol}
 (@var{window1} (@var{column1} . @var{row1})
  @var{buffer-pos1} @var{timestamp1})
 (@var{window2} (@var{column2} . @var{row2})
  @var{buffer-pos2} @var{timestamp2}))
@end example

The name of @var{button-symbol} contains the prefix @samp{drag-}.  The
second and third elements of the event give the starting and ending
position of the drag.

The @samp{drag-} prefix follows the modifier key prefixes such as
@samp{C-} and @samp{M-}.

If @code{read-key-sequence} receives a drag event which has no key
binding, and the corresponding click event does have a binding, it
changes the drag event into a click event at the drag's starting
position.  This means that you don't have to distinguish between click
and drag events unless you want to.

@item
Click and drag events happen when you release a mouse button.  Another
kind of event happens when you press a button.  It looks just like a
click event, except that the name of @var{button-symbol} contains the
prefix @samp{down-}.  The @samp{down-} prefix follows the modifier key
prefixes such as @samp{C-} and @samp{M-}.

The function @code{read-key-sequence}, and the Emacs command loop,
ignore any down events that don't have command bindings.  This means
that you need not worry about defining down events unless you want them
to do something.  The usual reason to define a down event is so that you
can track mouse motion until the button is released.

@item
For example, if the user presses and releases the left mouse button over
the same location, Emacs generates a sequence of events like this:

@smallexample
(down-mouse-1 (#<window 18 on NEWS> 2613 (0 . 38) -864320))
(mouse-1      (#<window 18 on NEWS> 2613 (0 . 38) -864180))
@end smallexample

Or, while holding the control key down, the user might hold down the
second mouse button, and drag the mouse from one line to the next.
That produces two events, as shown here:

@smallexample
(C-down-mouse-2 (#<window 18 on NEWS> 3440 (0 . 27) -731219))
(C-drag-mouse-2 (#<window 18 on NEWS> 3440 (0 . 27) -731219)
                (#<window 18 on NEWS> 3510 (0 . 28) -729648))
@end smallexample

Or, while holding down the meta and shift keys, the user might press
the second mouse button on the window's mode line, and then drag the
mouse into another window.  That produces an event like this:

@smallexample
(M-S-down-mouse-2 (#<window 18 on NEWS> mode-line (33 . 31) -457844))
(M-S-drag-mouse-2 (#<window 18 on NEWS> mode-line (33 . 31) -457844)
                  (#<window 20 on carlton-sanskrit.tex> 161 (33 . 3)
                   -453816))
@end smallexample

@item
A key sequence that starts with a mouse click is read using the keymaps
of the buffer in the window clicked on, not the current buffer.

This does not imply that clicking in a window selects that window or its
buffer.  The execution of the command begins with no change in the
selected window or current buffer.  However, the command can switch
windows or buffers if programmed to do so.

@item
Mouse motion events are represented by lists.  During the execution of
the body of a @code{track-mouse} form, moving the mouse generates events
that look like this:

@example
(mouse-movement (@var{window} (@var{column} . @var{row})
                 @var{buffer-pos} @var{timestamp}))
@end example

The second element of the list describes the current position of the
mouse, just as in a mouse click event.

Outside of @code{track-mouse} forms, Emacs does not generate events for
mere motion of the mouse, and these events do not appear.

@item
Focus shifts between frames are represented by lists.

When the mouse shifts temporary input focus from one frame to another,
Emacs generates an event like this:

@example
(switch-frame @var{new-frame})
@end example

@noindent
where @var{new-frame} is the frame switched to.

In X windows, most window managers are set up so that just moving the
mouse into a window is enough to set the focus there.  As far as the
user concern, Emacs behaves consistently with this.  However, there is
no need for the Lisp program to know about the focus change until some
other kind of input arrives.  So Emacs generates the focus event only
when the user actually types a keyboard key or presses a mouse button in
the new frame; just moving the mouse between frames does not generate a
focus event.

The global key map usually binds this event to the
@code{internal-select-frame} function, so that characters typed at a
frame apply to that frame's selected window.

If the user switches frames in the middle of a key sequence, then Emacs
delays the @code{switch-frame} event until the key sequence is over.
For example, suppose @kbd{C-c C-a} is a key sequence in the current
buffer's keymaps.  If the user types @kbd{C-c}, moves the mouse to
another frame, and then types @kbd{C-a}, @code{read-key-sequence}
returns the sequence @code{"\C-c\C-a"}, and the next call to
@code{read-event} or @code{read-key-sequence} will return the
@code{switch-frame} event.
@end itemize

@section Working with Input Events

@itemize @bullet
@item
Functions which work with key sequences now handle non-character
events.  Functions like @code{define-key}, @code{global-set-key}, and
@code{local-set-key} used to accept strings representing key sequences;
now, since events may be arbitrary lisp objects, they also accept
vectors.  The function @code{read-key-sequence} may return a string or a
vector, depending on whether or not the sequence read contains only
characters.

List events may be represented by the symbols at their head; to bind
clicks of the left mouse button, you need only present the symbol
@code{mouse-1}, not an entire mouse click event.  If you do put an event
which is a list in a key sequence, only the event's head symbol is used
in key lookups.

For example, to globally bind the left mouse button to the function
@code{mouse-set-point}, you could evaluate this:

@example
(global-set-key [mouse-1] 'mouse-set-point)
@end example

To bind the sequence @kbd{C-c @key{F1}} to the command @code{tex-view}
in @code{tex-mode-map}, you could evaluate this:

@example
(define-key tex-mode-map [?\C-c f1] 'tex-view)
@end example

To find the binding for the function key labeled @key{NEXT} in
@code{minibuffer-local-map}, you could evaluate this:

@example
(lookup-key minibuffer-local-map [next])
     @result{} next-history-element
@end example

If you call the function @code{read-key-sequence} and then press
@kbd{C-x C-@key{F5}}, here is how it behaves:

@example
(read-key-sequence "Press `C-x C-F5': ")
     @result{} [24 C-f5]
@end example

Note that @samp{24} is the character @kbd{C-x}.

@item
The documentation functions (@code{single-key-description},
@code{key-description}, etc.) now handle the new event types.  Wherever
a string of keyboard input characters was acceptable in previous
versions of Emacs, a vector of events should now work.

@item
Special parts of a window can have their own bindings for mouse events.

When mouse events occur in special parts of a window, such as a mode
line or a scroll bar, the event itself shows nothing special---only the
symbol that would normally represent that mouse button and modifier
keys.  The information about the screen region is kept in other parts
of the event list.  But @code{read-key-sequence} translates this
information into imaginary prefix keys, all of which are symbols:
@code{mode-line}, @code{vertical-line}, @code{horizontal-scrollbar} and
@code{vertical-scrollbar}.

For example, if you call @code{read-key-sequence} and then click the
mouse on the window's mode line, this is what happens:

@smallexample
(read-key-sequence "Click on the mode line: ")
     @result{} [mode-line (mouse-1 (#<window 6 on NEWS> mode-line
                              (40 . 63) 5959987))]
@end smallexample

You can define meanings for mouse clicks in special window regions by
defining key sequences using these imaginary prefix keys.  For example,
here is how to bind the third mouse button on a window's mode line
delete the window:

@example
(global-set-key [mode-line mouse-3] 'mouse-delete-window)
@end example

Here's how to bind the middle button (modified by @key{META}) on the
vertical line at the right of a window to scroll the window to the
left.

@example
(global-set-key [vertical-line M-mouse-2] 'scroll-left)
@end example

@item
Decomposing an event symbol.

Each symbol used to identify a function key or mouse button has a
property named @code{event-symbol-elements}, which is a list containing
an unmodified version of the symbol, followed by modifiers the symbol
name contains.  The modifiers are symbols; they include @code{shift},
@code{control}, and @code{meta}.  In addition, a mouse event symbol has
one of @code{click}, @code{drag}, and @code{down}.  For example:

@example
(get 'f5 'event-symbol-elements)
     @result{} (f5)
(get 'C-f5 'event-symbol-elements)
     @result{} (f5 control)
(get 'M-S-f5 'event-symbol-elements)
     @result{} (f5 meta shift)
(get 'mouse-1 'event-symbol-elements)
     @result{} (mouse-1 click)
(get 'down-mouse-1 'event-symbol-elements)
     @result{} (mouse-1 down)
@end example

Note that the @code{event-symbol-elements} property for a mouse click
explicitly contains @code{click}, but the event symbol name itself does
not contain @samp{click}.

@item
Use @code{read-event} to read input if you want to accept any kind of
event.  The old function @code{read-char} now discards events other than
keyboard characters.

@item
@code{last-command-char} and @code{last-input-char} can now hold any
kind of event.

@item
The new variable @code{unread-command-events} is much like
@code{unread-command-char}.  Its value is a list of events of any type,
to be processed as command input in order of appearance in the list.

@item
The function @code{this-command-keys} may return a string or a vector,
depending on whether or not the sequence read contains only characters.
You may need to upgrade code which uses this function.

The function @code{recent-keys} now returns a vector of events.
You may need to upgrade code which uses this function.

@item
A keyboard macro's definition can now be either a string or a vector.
All that really matters is what elements it has.  If the elements are
all characters, then the macro can be a string; otherwise, it has to be
a vector.

@item
The variable @code{last-event-frame} records which frame the last input
event was directed to.  Usually this is the frame that was selected when
the event was generated, but if that frame has redirected input focus to
another frame, @code{last-event-frame} is the frame to which the event
was redirected.

@item
The interactive specification now allows a new code letter @samp{e} to
simplify commands bound to events which are lists.  This code supplies
as an argument the complete event object.

You can use @samp{e} more than once in a single command's interactive
specification.  If the key sequence which invoked the command has
@var{n} events with parameters, the @var{n}th @samp{e} provides the
@var{n}th parameterized event.  Events which are not lists, such as
function keys and ASCII keystrokes, do not count where @samp{e} is
concerned.

@item
You can extract the starting and ending position values from a mouse
button or motion event using the two functions @code{event-start} and
@code{event-end}.  These two functions return different values for drag
and motion events; for click and button-down events, they both return
the position of the event.

@item
The position, a returned by @code{event-start} and @code{event-end}, is
a list of this form:

@example
(@var{window} @var{buffer-position} (@var{col} . @var{row}) @var{timestamp})
@end example

You can extract parts of this list with the functions
@code{posn-window}, @code{posn-point}, @code{posn-col-row}, and
@code{posn-timestamp}.

@item
The function @code{scroll-bar-scale} is useful for computing where to
scroll to in response to a mouse button event from a scroll bar.  It
takes two arguments, @var{ratio} and @var{total}, and in effect
multiplies them.  We say ``in effect'' because @var{ratio} is not a
number; rather a pair @code{(@var{num} . @var{denom}).

Here's the usual way to use @code{scroll-bar-scale}:

@example
(scroll-bar-scale (posn-col-row (event-start event))
                  (buffer-size))
@end example
@end itemize

@section Putting Keyboard Events in Strings

  In most of the places where strings are used, we conceptualize the
string as containing text characters---the same kind of characters found
in buffers or files.  Occasionally Lisp programs use strings which
conceptually contain keyboard characters; for example, they may be key
sequences or keyboard macro definitions.  There are special rules for
how to put keyboard characters into a string, because they are not
limited to the range of 0 to 255 as text characters are.

  A keyboard character typed using the @key{META} key is called a
@dfn{meta character}.  The numeric code for such an event includes the
2**23 bit; it does not even come close to fitting in a string.  However,
earlier Emacs versions used a different representation for these
characters, which gave them codes in the range of 128 to 255.  That did
fit in a string, and many Lisp programs contain string constants that
use @samp{\M-} to express meta characters, especially as the argument to
@code{define-key} and similar functions.

  We provide backward compatibility to run those programs with special
rules for how to put a keyboard character event in a string.  Here are
the rules:

@itemize @bullet
@item
If the keyboard event value is in the range of 0 to 127, it can go in the
string unchanged.

@item
The meta variants of those events, with codes in the range of 2**23 to
2**23+127, can also go in the string, but you must change their numeric
values.  You must set the 2**7 bit instead of the 2**23 bit, resulting
in a value between 128 and 255.

@item
Other keyboard character events cannot fit in a string.  This includes
keyboard events in the range of 128 to 255.
@end itemize

  Functions such as @code{read-key-sequence} that can construct strings
containing events follow these rules.

  When you use the read syntax @samp{\M-} in a string, it produces a
code in the range of 128 to 255---the same code that you get if you
modify the corresponding keyboard event to put it in the string.  Thus,
meta events in strings work consistently regardless of how they get into
the strings.

  New programs can avoid dealing with these rules by using vectors
instead of strings for key sequences when there is any possibility that
these issues might arise.

  The reason we changed the representation of meta characters as
keyboard events is to make room for basic character codes beyond 127,
and support meta variants of such larger character codes.

@section Menus

You can now define menus conveniently as keymaps.  Menus are normally
used with the mouse, but they can work with the keyboard also.

@subsection Defining Menus

A keymap is suitable for menu use if it has an @dfn{overall prompt
string}, which is a string that appears as an element of the keymap.  It
should describes the purpose of the menu.  The easiest way to construct
a keymap with a prompt string is to specify the string as an argument
when you run @code{make-keymap} or @code{make-sparse-keymap}.

The individual bindings in the menu keymap should also have prompt
strings; these strings are the items in the menu.  A binding with a
prompt string looks like this:

@example
(@var{char} @var{string} . @var{real-binding})
@end example

As far as @code{define-key} is concerned, the string is part of the
character's binding---the binding looks like this:

@example
(@var{string} . @var{real-binding}).
@end example

However, only @var{real-binding} is used for executing the key.

You can also supply a second string, called the help string, as follows:

@example
(@var{char} @var{string} @var{help-string} . @var{real-binding})
@end example

Currently Emacs does not actually use @var{help-string}; it knows only
how to ignore @var{help-string} in order to extract @var{real-binding}.
In the future we hope to make @var{help-string} serve as longer
documentation for the menu item, available on request.

The prompt string for a binding should be short---one or two words.  Its
meaning should describe the command it corresponds to.

If @var{real-binding} is @code{nil}, then @var{string} appears in the
menu but cannot be selected.

If @var{real-binding} is a symbol, and has a non-@code{nil}
@code{menu-enable} property, that property is an expression which
controls whether the menu item is enabled.  Every time the keymap is
used to display a menu, Emacs evaluates the expression, and it enables
the menu item only if the expression's value is non-@code{nil}.  When a
menu item is disabled, it is displayed in a ``fuzzy'' fashion, and
cannot be selected with the mouse.

@subsection Menus and the Mouse

The way to make a menu keymap produce a menu is to make it the
definition of a prefix key.

When the prefix key ends with a mouse event, Emacs handles the menu
keymap by popping up a visible menu that you can select from with the
mouse.  When you click on a menu item, the event generated is whatever
character or symbol has the binding which brought about that menu item.

A single keymap can appear as multiple panes, if you explicitly
arrange for this.  The way to do this is to make a keymap for each
pane, then create a binding for each of those maps in the main keymap
of the menu.  Give each of these bindings a prompt string that starts
with @samp{@@}.  The rest of the prompt string becomes the name of the
pane.  See the file @file{lisp/mouse.el} for an example of this.  Any
ordinary bindings with prompt strings are grouped into one pane, which
appears along with the other panes explicitly created for the
submaps.

You can also get multiple panes from separate keymaps.  The full
definition of a prefix key always comes from merging the definitions
supplied by the various active keymaps (minor modes, local, and
global).  When more than one of these keymaps is a menu, each of them
makes a separate pane or panes.

@subsection Menus and the Keyboard

When a prefix key ending with a keyboard event (a character or function
key) has a definition that is a menu keymap, you can use the keyboard
to choose a menu item.

Emacs displays the menu alternatives in the echo area.  If they don't
all fit at once, type @key{SPC} to see the next line of alternatives.
If you keep typing @key{SPC}, you eventually get to the end of the menu
and then cycle around to the beginning again.

When you have found the alternative you want, type the corresponding
character---the one whose binding is that alternative.

In a menu intended for keyboard use, each menu item must clearly
indicate what character to type.  The best convention to use is to make
the character the first letter of the menu item prompt string.  That is
something users will understand without being told.

@subsection The Menu Bar

  Under X Windows, each frame can have a @dfn{menu bar}---a permanently
displayed menu stretching horizontally across the top of the frame.  The
items of the menu bar are the subcommands of the fake ``function key''
@code{menu-bar}, as defined by all the active keymaps.

  To add an item to the menu bar, invent a fake ``function key'' of your
own (let's call it @var{key}), and make a binding for the key sequence
@code{[menu-bar @var{key}]}.  Most often, the binding is a menu keymap,
so that pressing a button on the menu bar item leads to another menu.

  In order for a frame to display a menu bar, its @code{menu-bar-lines}
property must be greater than zero.  Emacs uses just one line for the
menu bar itself; if you specify more than one line, the other lines
serve to separate the menu bar from the windows in the frame.  We
recommend you try one or two as the @code{menu-bar-lines} value.

@section Keymaps

@itemize @bullet
@item
The representation of keymaps has changed to support the new event
types.  All keymaps now have the form @code{(keymap @var{element}
@var{element} @dots{})}.  Each @var{element} takes one of the following
forms:

@table @asis
@item @var{prompt-string}
A string as an element of the keymap marks the keymap as a menu, and
serves as the overal prompt string for it.

@item @code{(@var{key} . @var{binding})}
A cons cell binds @var{key} to @var{definition}.  Here @var{key} may be
any sort of event head---a character, a function key symbol, or a mouse
button symbol.

@item @var{vector}
A vector of 128 elements binds all the ASCII characters; the @var{n}th
element holds the binding for character number @var{n}.

@item @code{(t . @var{binding})}
A cons cell whose @sc{car} is @code{t} is a default binding; anything
not bound by previous keymap elements is given @var{binding} as its
binding.

Default bindings are important because they allow a keymap to bind all
possible events without having to enumerate all the possible function
keys and mouse clicks, with all possible modifier prefixes.

The function @code{lookup-key} (and likewise other functions for
examining a key binding) normally report only explicit bindings of the
specified key sequence; if there is none, they return @code{nil}, even
if there is a default binding that would apply to that key sequence if
it were actually typed in.  However, these functions now take an
optional argument @var{accept-defaults} which, if non-@code{nil}, says
to consider default bindings.

Note that if a vector in the keymap binds an ASCII character to
@code{nil} (thus making it ``unbound''), the default binding does not
apply to the character.  Think of the vector element as an explicit
binding of @code{nil}.

Note also that if the keymap for a minor or major mode contains a
default binding, it completely masks out any lower-priority keymaps.
@end table

@item
A keymap can now inherit from another keymap.  Do do this, make the
latter keymap the ``tail'' of the new one.  Such a keymap looks like
this:

@example
(keymap @var{bindings}@dots{} . @var{other-keymap})
@end example

The effect is that this keymap inherits all the bindings of
@var{other-keymap}, but can add to them or override them with
@var{bindings}.  Subsequent changes in the bindings of
@var{other-keymap} @emph{do} affect this keymap.

For example,

@example
(setq my-mode-map (cons 'keymap text-mode-map))
@end example

@noindent
makes a keymap that by default inherits all the bindings of Text
mode---whatever they may be at the time a key is looked up.  Any
bindings made explicitly in @code{my-mode-map} override the bindings
inherited from Text mode, however.

@item
Minor modes can now have local keymaps.  Thus, a key can act a special
way when a minor mode is in effect, and then revert to the major mode or
global definition when the minor mode is no longer in effect.  The
precedence of keymaps is now: minor modes (in no particular order), then
major mode, and lastly the global map.

The new @code{current-minor-mode-maps} function returns a list of all
the keymaps of currently enabled minor modes, in the other that they
apply.

To set up a keymap for a minor mode, add an element to the alist
@code{minor-mode-map-alist}.  Its elements look like this:

@example
(@var{symbol} . @var{keymap})
@end example

The keymap @var{keymap} is active whenever @var{symbol} has a
non-@code{nil} value.  Use for @var{symbol} the variable which indicates
whether the minor mode is enabled.

When more than one minor mode keymap is active, their order of
precedence is the order of @code{minor-mode-map-alist}.  But you should
design minor modes so that they don't interfere with each other, and if
you do this properly, the order will not matter.

The function @code{minor-mode-key-binding} returns a list of all the
active minor mode bindings of @var{key}.  More precisely, it returns an
alist of pairs @code{(@var{modename} . @var{binding})}, where
@var{modename} is the the variable which enables the minor mode, and
@var{binding} is @var{key}'s definition in that mode.  If @var{key} has
no minor-mode bindings, the value is @code{nil}.

If the first binding is a non-prefix, all subsequent bindings from other
minor modes are omitted, since they would be completely shadowed.
Similarly, the list omits non-prefix bindings that follow prefix
bindings.

@item
The new function @code{copy-keymap} copies a keymap, producing a new
keymap with the same key bindings in it.  If the keymap contains other
keymaps directly, these subkeymaps are copied recursively.

If you want to, you can define a prefix key with a binding that is a
symbol whose function definition is another keymap.  In this case,
@code{copy-keymap} does not look past the symbol; it doesn't copy the
keymap inside the symbol.

@item
@code{substitute-key-definition} now accepts an optional fourth
argument, which is a keymap to use as a template.

@example
(substitute-key-definition olddef newdef keymap oldmap)
@end example

@noindent
finds all characters defined in @var{oldmap} as @var{olddef},
and defines them in @var{keymap} as @var{newdef}.

In addition, this function now operates recursively on the keymaps that
define prefix keys within @var{keymap} and @var{oldmap}.
@end itemize

@section Minibuffer Features

The minibuffer input functions @code{read-from-minibuffer} and
@code{completing-read} have new features.

@subsection Minibuffer History

A new optional argument @var{hist} specifies which history list to use.
If you specify a variable (a symbol), that variable is the history
list.  If you specify a cons cell @code{(@var{variable}
. @var{startpos})}, then @var{variable} is the history list variable,
and @var{startpos} specifies the initial history position (an integer,
counting from zero which specifies the most recent element of the
history).

If you specify @var{startpos}, then you should also specify that element
of the history as @var{initial-input}, for consistency.

If you don't specify @var{hist}, then the default history list
@code{minibuffer-history} is used.  Other standard history lists that
you can use when appropriate include @code{query-replace-history},
@code{command-history}, and @code{file-name-history}.

The value of the history list variable is a list of strings, most recent
first.  You should set a history list variable to @code{nil} before
using it for the first time.

@code{read-from-minibuffer} and @code{completing-read} add new elements
to the history list automatically, and provide commands to allow the
user to reuse items on the list.  The only thing your program needs to
do to use a history list is to initialize it and to pass its name to the
input functions when you wish.  But it is safe to modify the list by
hand when the minibuffer input functions are not using it.

@subsection Other Minibuffer Features

The @var{initial} argument to @code{read-from-minibufer} and other
minibuffer input functions can now be a cons cell @code{(@var{string}
. @var{position})}.  This means to start off with @var{string} in the
minibuffer, but put the cursor @var{position} characters from the
beginning, rather than at the end.

In @code{read-no-blanks-input}, the @var{initial} argument is now
optional; if it is omitted, the initial input string is the empty
string.

@section New Features for Defining Commands

@itemize @bullet
@item
If the interactive specification begins with @samp{@@}, this means to
select the window under the mouse.  This selection takes place before
doing anything else with the command.

You can use both @samp{@@} and @samp{*} together in one command; they
are processed in order of appearance.

@item
Prompts in an interactive specification can incorporate the values of
the preceding arguments.  Emacs replaces @samp{%}-sequences (as used
with the @code{format} function) in the prompt with the interactive
arguments that have been read so far.  For example, a command with this
interactive specification

@example
(interactive "sReplace: \nsReplace %s with: ")
@end example

@noindent
prompts for the first argument with @samp{Replace: }, and then prompts
for the second argument with @samp{Replace @var{foo} with: }, where
@var{foo} is the string read as the first argument.

@item
If a command name has a property @code{enable-recursive-minibuffers}
which is non-@code{nil}, then the command can use the minibuffer to read
arguments even if it is invoked from the minibuffer.  The minibuffer
command @code{next-matching-history-element} (normally bound to
@kbd{M-s} in the minibuffer) uses this feature.
@end itemize

@section New Features for Reading Input

@itemize @bullet
@item
The function @code{set-input-mode} now takes four arguments.  The last
argument is optional.  Their names are @var{interrupt}, @var{flow},
@var{meta} and @var{quit}.

The argument @var{interrupt} says whether to use interrupt-driven
input.  Non-@code{nil} means yes, and @code{nil} means no (use CBREAK
mode).

The argument @var{flow} says whether to enable terminal flow control.
Non-@code{nil} means yes.

The argument @var{meta} says whether to enable the use of a Meta key.
Non-@code{nil} means yes.

If @var{quit} non-@code{nil}, it is the character to use for quitting.
(Normally this is @kbd{C-g}.)

@item
The variable @code{meta-flag} has been deleted; use
@code{set-input-mode} to enable or disable support for a @key{META}
key.  This change was made because @code{set-input-mode} can send the
terminal the appropriate commands to enable or disable operation of the
@key{META} key.

@item
The new variable @code{extra-keyboard-modifiers} lets Lisp programs
``press'' the modifier keys on the keyboard.
The value is a bit mask:

@table @asis
@item 1
The @key{SHIFT} key.
@item 2
The @key{LOCK} key.
@item 4
The @key{CTL} key.
@item 8
The @key{META} key.
@end table

When you use X windows, the program can press any of the modifier keys
in this way.  Otherwise, only the @key{CTL} and @key{META} keys can be
virtually pressed.

@item
You can use the new function @code{keyboard-translate} to set up
@code{keyboard-translate-table} conveniently.

@item
Y-or-n questions using the @code{y-or-n-p} function now accept @kbd{C-]}
(usually mapped to @code{abort-recursive-edit}) as well as @kbd{C-g} to
quit.

@item
The variable @code{num-input-keys} is the total number of key sequences
that the user has typed during this Emacs session.

@item
A new Lisp variable, @code{function-key-map}, holds a keymap which
describes the character sequences sent by function keys on an ordinary
character terminal.  This uses the same keymap data structure that is
used to hold bindings of key sequences, but it has a different meaning:
it specifies translations to make while reading a key sequence.

If @code{function-key-map} ``binds'' a key sequence @var{k} to a vector
@var{v}, then when @var{k} appears as a subsequence @emph{anywhere} in a
key sequence, it is replaced with @var{v}.

For example, VT100 terminals send @kbd{@key{ESC} O P} when the ``keypad''
PF1 key is pressed.  Thus, on a VT100, @code{function-key-map} should
``bind'' that sequence to @code{[pf1]}.  This specifies translation of
@kbd{@key{ESC} O P} into @key{PF1} anywhere in a key sequence.

Thus, typing @kbd{C-c @key{PF1}} sends the character sequence @kbd{C-c
@key{ESC} O P}, but @code{read-key-sequence} translates this back into
@kbd{C-c @key{PF1}}, which it returns as the vector @code{[?\C-c PF1]}.

Entries in @code{function-key-map} are ignored if they conflict with
bindings made in the minor mode, local, or global keymaps.

The value of @code{function-key-map} is usually set up automatically
according to the terminal's Terminfo or Termcap entry, and the
terminal-specific Lisp files.  Emacs comes with a number of
terminal-specific files for many common terminals; their main purpose is
to make entries in @code{function-key-map} beyond those that can be
deduced from Termcap and Terminfo.

@item
The variable @code{key-translation-map} works like @code{function-key-map}
except for two things:

@itemize @bullet
@item
@code{key-translation-map} goes to work after @code{function-key-map} is
finished; it receives the results of translation by
@code{function-key-map}.

@item
@code{key-translation-map} overrides actual key bindings.
@end itemize

The intent of @code{key-translation-map} is for users to map one
character set to another, including ordinary characters normally bound
to @code{self-insert-command}.
@end itemize

@section New Syntax Table Features

@itemize @bullet
@item
You can use two new functions to move across characters in certain
syntax classes.

@code{skip-syntax-forward} moves point forward across characters whose
syntax classes are mentioned in its first argument, a string.  It stops
when it encounters the end of the buffer, or position @var{lim} (the
optional second argument), or a character it is not supposed to skip.
The function @code{skip-syntax-backward} is similar but moves backward.

@item
The new function @code{forward-comment} moves point by comments.  It
takes one argument, @var{count}; it moves point forward across
@var{count} comments (backward, if @var{count} is negative).  If it
finds anything other than a comment or whitespace, it stops, leaving
point at the far side of the last comment found.  It also stops after
satisfying @var{count}.

@item
The new variable @code{words-include-escapes} affects the behavior of
@code{forward-word} and everything that uses it.  If it is
non-@code{nil}, then characters in the ``escape'' and ``character
quote'' syntax classes count as part of words.

@item
There are two new syntax flags for use in syntax tables.

@itemize -
@item
The prefix flag.

The @samp{p} flag identifies additional ``prefix characters'' in Lisp
syntax.  You can set this flag with @code{modify-syntax-entry} by
including the letter @samp{p} in the syntax specification.

These characters are treated as whitespace when they appear between
expressions.  When they appear withing an expression, they are handled
according to their usual syntax codes.

The function @code{backward-prefix-chars} moves back over these
characters, as well as over characters whose primary syntax class is
prefix (@samp{'}).

@item
The @samp{b} comment style flag.

Emacs can now supports two comment styles simultaneously.  (This is for
the sake of C++.)  More specifically, it can recognize two different
comment-start sequences.  Both must share the same first character; only
the second character may differ.  Mark the second character of the
@samp{b}-style comment start sequence with the @samp{b} flag.  You can
set this flag with @code{modify-syntax-entry} by including the letter
@samp{b} in the syntax specification.

The two styles of comment can have different comment-end sequences.  A
comment-end sequence (one or two characters) applies to the @samp{b}
style if its first character has the @samp{b} flag set; otherwise, it
applies to the @samp{a} style.

The appropriate comment syntax settings for C++ are as follows:

@table @asis
@item @samp{/}
@samp{124b}
@item @samp{*}
@samp{23}
@item newline
@samp{>b}
@end table

Thus @samp{/*} is a comment-start sequence for @samp{a} style, @samp{//}
is a comment-start sequence for @samp{b} style, @samp{*/} is a
comment-end sequence for @samp{a} style, and newline is a comment-end
sequence for @samp{b} style.
@end itemize
@end itemize

@section The Case Table

You can customize case conversion using the new case table feature.  A
case table is a collection of strings that specifies the mapping between
upper case and lower case letters.  Each buffer has its own case table.
You need a case table if you are using a language which has letters that
are not standard ASCII letters.

A case table is a list of this form:

@example
(@var{downcase} @var{upcase} @var{canonicalize} @var{equivalences})
@end example

@noindent
where each element is either @code{nil} or a string of length 256.  The
element @var{downcase} says how to map each character to its lower-case
equivalent.  The element @var{upcase} maps each character to its
upper-case equivalent.  If lower and upper case characters are in 1-1
correspondence, use @code{nil} for @var{upcase}; then Emacs deduces the
upcase table from @var{downcase}.

For some languages, upper and lower case letters are not in 1-1
correspondence.  There may be two different lower case letters with the
same upper case equivalent.  In these cases, you need to specify the
maps for both directions.

The element @var{canonicalize} maps each character to a canonical
equivalent; any two characters that are related by case-conversion have
the same canonical equivalent character.

The element @var{equivalences} is a map that cyclicly permutes each
equivalence class (of characters with the same canonical equivalent).

You can provide @code{nil} for both @var{canonicalize} and
@var{equivalences}, in which case both are deduced from @var{downcase}
and @var{upcase}.

Here are the functions for working with case tables:

@code{case-table-p} is a predicate that says whether a Lisp object is a
valid case table.

@code{set-standard-case-table} takes one argument and makes that
argument the case table for new buffers created subsequently.
@code{standard-case-table} returns the current value of the new buffer
case table.

@code{current-case-table} returns the case table of the current buffer.
@code{set-case-table} sets the current buffer's case table to the
argument.

@code{set-case-syntax-pair} is a convenient function for specifying a
pair of letters, upper case and lower case.  Call it with two arguments,
the upper case letter and the lower case letter.  It modifies the
standard case table and a few syntax tables that are predefined in
Emacs.  This function is intended as a subroutine for packages that
define non-ASCII character sets.

Load the library @file{iso-syntax} to set up the syntax and case table for
the 256 bit ISO Latin 1 character set.

@section New Features for Dealing with Buffers

@itemize @bullet
@item
The new function @code{buffer-modified-tick} returns a buffer's
modification-count that ticks every time the buffer is modified.  It
takes one optional argument, which is the buffer you want to examine.
If the argument is @code{nil} (or omitted), the current buffer is used.

@item
@code{buffer-disable-undo} is a new name for the function
formerly known as @code{buffer-flush-undo}.  This turns off recording
of undo information in the buffer given as argument.

@item
The new function @code{generate-new-buffer-name} chooses a name that
would be unique for a new buffer---but does not create the buffer.  Give
it one argument, a starting name.  It produces a name not in use for a
buffer by appending a number inside of @samp{<@dots{}>}.

@item
The function @code{rename-buffer} now takes an option second argument
which tells it that if the specified new name corresponds to an existing
buffer, it should use @code{generate-new-buffer-name} to modify the name
to be unique, rather than signaling an error.

@code{rename-buffer} now returns the name to which the buffer was
renamed.

@item
The function @code{list-buffers} now looks at the local variable
@code{list-buffers-directory} in each non-file-visiting buffer, and
shows its value where the file would normally go.  Dired sets this
variable in each Dired buffer, so the buffer list now shows which
directory each Dired buffer is editing.

@item
The function @code{other-buffer} now takes an optional second argument
@var{visible-ok} which, if non-@code{nil}, indicates that buffers
currently being displayed in windows may be returned even if there are
other buffers not visible.  Normally, @code{other-buffer} returns a
currently visible buffer only as a last resort, if there are no suitable
nonvisible buffers.

@item
The hook @code{kill-buffer-hook} now runs whenever a buffer is killed.
@end itemize

@section Local Variables Features

@itemize @bullet
@item
If a local variable name has a non-@code{nil} @code{permanent-local}
property, then @code{kill-all-local-variables} does not kill it.  Such
local variables are ``permanent''---they remain unchanged even if you
select a different major mode.

Permanent locals are useful when they have to do with where the file
came from or how to save it, rather than with how to edit the contents.

@item
The function @code{make-local-variable} now never changes the value of the variable
that it makes local.  If the variable had no value before, it still has
no value after becoming local.

@item
The new function @code{default-boundp} tells you whether a variable has
a default value (as opposed to being unbound in its default value).  If
@code{(default-boundp 'foo)} returns @code{nil}, then
@code{(default-value 'foo)} would get an error.

@code{default-boundp} is to @code{default-value} as @code{boundp} is to
@code{symbol-value}.

@item
The special forms @code{defconst} and @code{defvar}, when the variable
is local in the current buffer, now set the variable's default value
rather than its local value.
@end itemize

@section New Features for Subprocesses

@itemize @bullet
@item
@code{call-process} and @code{call-process-region} now return a value
that indicates how the synchronous subprocess terminated.  It is either
a number, which is the exit status of a process, or a signal name
represented as a string.

@item
@code{process-status} now returns @code{open} and @code{closed} as the
status values for network connections.

@item
The standard asynchronous subprocess features work on VMS now,
and the special VMS asynchronous subprocess functions have been deleted.

@item
You can use the transaction queue feature for more convenient
communication with subprocesses using transactions.

Call @code{tq-create} to create a transaction queue communicating with a
specified process.  Then you can call @code{tq-enqueue} to send a
transaction.  @code{tq-enqueue} takes these five arguments:

@example
(tq-enqueue @var{tq} @var{question} @var{regexp} @var{closure} @var{fn})
@end example

@var{tq} is the queue to use.  (Specifying the queue has the effect of
specifying the process to talk to.)  The argument @var{question} is the
outgoing message which starts the transaction.  The argument @var{fn} is
the function to call when the corresponding answer comes back; it is
called with two arguments: @var{closure}, and the answer received.

The argument @var{regexp} is a regular expression to match the entire
answer; that's how @code{tq-enqueue} tells where the answer ends.

Call @code{tq-close} to shut down a transaction queue and terminate its
subprocess.

@item
The function @code{signal-process} sends a signal to process @var{pid},
which need not be a child of Emacs.  The second argument @var{signal}
specifies which signal to send; it should be an integer.
@end itemize

@section New Features for Dealing with Times And Time Delays

@itemize @bullet
@item
The new function @code{current-time} returns the system's time value as
a list of three integers: @code{(@var{high} @var{low} @var{microsec})}.
The integers @var{high} and @var{low} combine to give the number of
seconds since 0:00 January 1, 1970, which is @var{high} * 2**16 +
@var{low}.

@var{microsec} gives the microseconds since the start of the current
second (or 0 for systems that return time only on the resolution of a
second).

@item
The function @code{current-time-string} accepts an optional argument
@var{time-value}.  If given, this specifies a time to format instead of
the current time.  The argument should be a cons cell containing two
integers, or a list whose first two elements are integers.  Thus, you
can use times obtained from @code{current-time} (see above) and from
@code{file-attributes}.

@item
You can now find out the user's time zone using @code{current-time-zone}.
It takes no arguments, and returns a list of this form:

@example
(@var{offset} @var{savings-flag} @var{standard} @var{savings})
@end example

@var{offset} is an integer specifying how many minutes east of Greenwich
the current time zone is located.  A negative value means west of
Greenwich.  Note that this describes the standard time; if daylight
savings time is in effect, it does not affect this value.

@var{savings-flag} is non-@code{nil} iff daylight savings time or some other
sort of seasonal time adjustment is in effect.

@var{standard} is a string giving the name of the time zone when no
seasonal time adjustment is in effect.

@var{savings} is a string giving the name of the time zone when there is a
seasonal time adjustment in effect.

If the user has specified a region that does not use a seasonal time
adjustment, @var{savings-flag} is always @code{nil}, and @var{standard}
and @var{savings} are equal.

@item
@code{sit-for}, @code{sleep-for} now let you specify the time period in
milliseconds as well as in seconds.  The first argument gives the number
of seconds, as before, and the optional second argument gives additional
milliseconds.  The time periods specified by these two arguments are
added together.

Not all systems support this; you get an error if you specify nonzero
milliseconds and it isn't supported.

@code{sit-for} also accepts an optional third argument @var{nodisp}.  If
this is non-@code{nil}, @code{sit-for} does not redisplay.  It still
waits for the specified time or until input is available.

@item
@code{accept-process-output} now accepts a timeout specified by optional
second and third arguments.  The second argument specifies the number of
seconds, while the third specifies the number of milliseconds.  The time
periods specified by these two arguments are added together.

Not all systems support this; you get an error if you specify nonzero
milliseconds and it isn't supported.

The function returns @code{nil} if the timeout expired before output
arrived, or non-@code{nil} if it did get some output.

@item
You can set up a timer to call a function at a specified future time.
To do so, call @code{run-at-time}, like this:

@example
(run-at-time @var{time} @var{repeat} @var{function} @var{args}@dots{})
@end example

Here, @var{time} is a string saying when to call the function.  The
argument @var{function} is the function to call later, and @var{args}
are the arguments to give it when it is called.

The argument @var{repeat} specifies how often to repeat the call.  If
@var{repeat} is @code{nil}, there are no repetitions; @var{function} is
called just once, at @var{time}.  If @var{repeat} is an integer, it
specifies a repetition period measured in seconds.

Absolute times may be specified in a wide variety of formats; The form
@samp{@var{hour}:@var{min}:@var{sec} @var{timezone}
@var{month}/@var{day}/@var{year}}, where all fields are numbers, works;
the format that @code{current-time-string} returns is also allowed.

To specify a relative time, use numbers followed by units.
For example:

@table @samp
@item 1 min
denotes 1 minute from now.
@item 1 min 5 sec
denotes 65 seconds from now.
@item 1 min 2 sec 3 hour 4 day 5 week 6 fortnight 7 month 8 year
denotes exactly 103 months, 123 days, and 10862 seconds from now.
@end table

If @var{time} is an integer, that specifies a relative time measured in
seconds.
@end itemize

To cancel the requested future action, pass the value that @code{run-at-time}
returned to the function @code{cancel-timer}.

@section Profiling Lisp Programs

You can now make execution-time profiles of Emacs Lisp programs using
the @file{profile} library.  See the file @file{profile.el} for
instructions; if you have written a Lisp program big enough to be worth
profiling, you can surely understand them.

@section New Features for Lisp Debuggers

@itemize @bullet
@item
You can now specify which kinds of errors should invoke the Lisp
debugger by setting the variable @code{debug-on-error} to a list of error
conditions.  For example, if you set it to the list @code{(void-variable)},
then only errors about a variable that has no value invoke the
debugger.

@item
The variable @code{command-debug-status} is used by Lisp debuggers.  It
records the debugging status of current interactive command.  Each time
a command is called interactively, this variable is bound to
@code{nil}.  The debugger can set this variable to leave information for
future debugger invocations during the same command.

The advantage of this variable over some other variable in the debugger
itself is that the data will not be visible for any other command
invocation.

@item
The function @code{backtrace-frame} is intended for use in Lisp
debuggers.  It returns information about what a frame on the Lisp call
stack is doing.  You specify one argument, which is the number of stack
frames to count up from the current execution point.

If that stack frame has not evaluated the arguments yet (or is a special
form), the value is @code{(nil @var{function} @var{arg-forms}@dots{})}.

If that stack frame has evaluated its arguments and called its function
already, the value is @code{(t @var{function}
@var{arg-values}@dots{})}.

In the return value, @var{function} is whatever was supplied as @sc{car}
of evaluated list, or a @code{lambda} expression in the case of a macro
call.  If the function has a @code{&rest} argument, that is represented
as the tail of the list @var{arg-values}.

If the argument is out of range, @code{backtrace-frame} returns
@code{nil}.
@end itemize

@ignore

@item
@code{kill-ring-save} now gives visual feedback to indicate the region
of text being added to the kill ring.  If the opposite end of the
region is visible in the current window, the cursor blinks there.
Otherwise, some text from the other end of the region is displayed in
the message area.
@end ignore

@section Memory Allocation Changes

The list that @code{garbage-collect} returns now has one additional
element.  This is a cons cell containing two numbers.  It gives
information about the number of used and free floating point numbers,
much as the first element gives such information about the number of
used and free cons cells.

The new function @code{memory-limit} returns an indication of the last
address allocated by Emacs.  More precisely, it returns that address
divided by 1024.  You can use this to get a general idea of how your
actions affect the memory usage.

@section Hook Changes

@itemize @bullet
@item
Expanding an abbrev first runs the new hook
@code{pre-abbrev-expand-hook}.

@item
The editor command loop runs the normal hook @code{pre-command-hook}
before each command, and runs @code{post-command-hook} after each
command.

@item
Auto-saving runs the new hook @code{auto-save-hook} before actually
starting to save any files.

@item
The new variable @code{revert-buffer-insert-file-contents-function}
holds a function that @code{revert-buffer} now uses to read in the
contents of the reverted buffer---instead of calling
@code{insert-file-contents}.

@item
The variable @code{lisp-indent-hook} has been renamed to
@code{lisp-indent-function}.

@item
The variable @code{auto-fill-hook} has been renamed to
@code{auto-fill-function}.

@item
The variable @code{blink-paren-hook} has been renamed to
@code{blink-paren-function}.

@item
The variable @code{temp-buffer-show-hook} has been renamed to
@code{temp-buffer-show-function}.

@item
The variable @code{suspend-hook} has been renamed to
@code{suspend-hooks}, because it is a list of functions but is not a
normal hook.

@item
The new function @code{add-hook} provides a handy way to add a function
to a hook variable.  For example,

@example
(add-hook 'text-mode-hook 'my-text-hook-function)
@end example

@noindent
arranges to call @code{my-text-hook-function}
when entering Text mode or related modes.
@end itemize

@bye
author	Kenichi Handa <handa@m17n.org>
date	Fri, 04 Aug 2000 02:26:32 +0000
parents	58b646d9fefc
children