Commit 0b35fc88 authored by David Byers's avatar David Byers
Browse files

Added some documentation on variables, utility functions and local versions

of regular Emacs functions.
parent 7595c625
David Byers
This file lists utility functions and functions defined in the
LysKOM code that replace standard functions. Always use the
LysKOM-specific function when available.
All developers need to know this.
Functions that are replaced by lyskom versions
nbutlast * lyskom-nbutlast
butlast * lyskom-butlast
looking-at * lyskom-looking-at
truncate-string-to-width * lyskom-truncate-string-to-width
princ * lyskom-princ [1]
lookup-key * lyskom-lookup-key
set-default lyskom-set-default
setq-default lyskom-setq-default
default-value lyskom-default-value
default-value lyskom-default-value-safe
run-hook-with-args lyskom-run-hook-with-args
add-hook lyskom-add-hook
remove-hook lyskom-remove-hook
insert * lyskom-insert [2]
* lyskom-insert-at-point [2]
* lyskom-insert-before-prompt [2]
format * lyskom-format [2]
* lyskom-format-insert [2]
* lyskom-format-insert-at-point [2]
* lyskom-format-insert-before-prompt [2]
message * lyskom-message [2]
error * lyskom-error [2]
fill-region * lyskom-fill-region
beep * lyskom-beep [2]
ding * lyskom-beep [2]
read-string * lyskom-read-string [2]
* lyskom-read-number [2]
y-or-n-p * j-or-n-p [2]
yes-or-no-p * ja-or-nej-p [2]
mapc lyskom-traverse
mapcar lyskom-traverse
use-local-map * lyskom-use-local-map
ignore-errors * lyskom-ignore-errors
save-excursion lyskom-save-excursion
generate-new-buffer * lyskom-generate-new-buffer [2]
get-buffer-create * lyskom-get-buffer-create [2]
display-buffer * lyskom-display-buffer [2]
delete-window * lyskom-undisplay-buffer [2]
defvar def-kom-var
defun def-kom-command
defstruct def-komtype
read-from-minibuffer lyskom-with-lyskom-minibuffer
completing-read lyskom-with-lyskom-minibuffer
Note that in some cases "replaced" isn't exactly the right word.
the LysKOM functions may have different semantics, but if you're
thinking about one of the functions in the left-hand column, you
should have a look at the function in the right-hand column. The
functions that are straight replacements are marked with an
[1] Restrictions on the type of stream. Only does special stuff
for strings. If you're printing other objects you might as
well use princ.
[2] Perhaps not straight replacements, but they should always be
used in place of the non-lyskom functions.
Creating lists from other lists (and other sources)
Convert a vector to a list.
Return a copy of the N first elements of a list.
Like standard Common Lisp nbutlast. Removes the last N
elements from a list.
Like standard Common Lisp butlast. Returns a copy of a list
with the last N elements removed.
Return a pointer to the first list element that is not a zero.
For example (skip-first-zeroes '(0 0 1 0 3)) returns (1 0 3).
Return a copy of a list containing only those elements that
pass a particular test. The test function is a one-argument
function that returns non-nil if the argument passes. For
example (filter-list 'symbolp '(1 2 a b c) returns (a b c).
This function is currently recursive, so it cannot handle very
long lists.
Like mapcar, but iterates over two lists for as long as both
have elements (thus skipping the last elements of the longer
list). The function argument should be a function that takes
two arguments, the first an element from the first list and
the second an element from the other list. The function
returns a list of the result. For example (mapcar2 'cons '(a b
c d) '(1 2 3)) returns ((a . 1) (b . 2) (c . 3)).
Return the intersection of two lists.
Check if a list is a subset of another list.
The set functions are not designed to be efficient at all, so be
careful not to feed them huge amounts of data. They aren't
recursive though, so they shouldn't run out of stack space.
Manipulating Lists
Destructively insert a value in a list before the first
element eq to a particular value.
Destructively rotate a list so that a specific element becomes
the first element of the list. For example (lyskom-rotate-list
'(a b c d) 'c) will return '(c d a b).
Destructively move an element from one position in a list to
another. The first element eq to a particular value is moved
so that it will be located at the specified position in the
list. For example (lyskom-move-in-list '(a b c d) 'b 3) will
return (a c d b).
Note that all destructive functions may change the first element
of a list (or move it around), so if you want to change the value
of a particular variable using these functions you need to use
setq, or the results may not be what you expect.
Finding things in lists
Do assoc using the last element of each alist element as the
key. This function is pretty slow and creates lots of new
short-lived cons cells. Use with care. Note that this is *not*
the same as rassoc, which compares the cdr of list elements.
This function actually compares the *last* element of list
Return the cons cell of a list immediately preceding the first
cons cell in the list that has a car eq to a particular value.
This is useful if you need the element before a particular
element in a list. For example (lyskom-preceding-cons '(a b c
d) 'c) will return (b c d).
Like assoc, but does case-insensitive string comparisons of
the key (using downcase, not lyskom-unicase).
Like rassoc, but does case-insensitive string comparisons of
the key (using downcase, not lyskom-unicase).
Traverse any sequence, applying a function to each element in
turn. The value returned is always nil. This is what you want
when you're thinking about mapc or mapcar (and don't want the
Stop traversal started with lyskom-traverse. This allows you
to stop traversal prematurely. The rest of the body is still
executed, so it's not like break in C or C++. It just makes
sure that there are no more iterations.
Type and range checking
Return the maximum integer value the system can represent. If
your Emacs happens to implement bignums for integers you will
want to rewrite this function...
Check if the argument is a valid regexp. This function
attempts to use the argument as a regexp and catches the error
if it isn't one.
Compiler warnings
Ignore all arguments. This kills compiler warnings about
variables being bound but not used. This should only be used
in those cases where a variable is bound so it has a dynamic
binding when calling some other function. Use with care.
Specify that a function is loaded from elsewhere. This will
stop the compiler from complaining about it being undefined.
Loading and using external functions
Attempt to load a library using require, but don't die if it
is unavailable. Optional arguments can specify an error
message to insert in the LysKOM buffer if the library cannot
be loaded.
Implementing compatibility between Emacs versions
Return a symbol representing the Emacs version. The result is
either emacs or xemacs. You should never need this. Use
lyskom-xemacs-or-gnu instead.
Process input and output
Like accept-process-output but behaves nicely on Emacs
versions with broken process handling (like XEmacs). This
function will call accept-process-output with increasing
timeouts (up to a maximum) until process output is avaiable.
Call lyskom-reset-apo-timeout after reading process output to
reset the timeout.
Date and time functions
Convert the current time in a format that LysKOM understands.
The result can be used as the argument list to
lyskom-create-time. Use this if you want to put the local time
into a LysKOM data structure that contains a LysKOM time.
Return a string containg the current time and date.
This appears to be very similar to lyskom-current-time.
String encoding and decoding, aka MULE, aka Pain in the Ass
If multibyte characters are not supported, encode the string
using the default coding system for the language or raw text.
Does nothing if multibyte characters are supported.
String and character manipulation
Convert a character to the canonical case for the character
according the current collate table (which should have been
retreived from the server).
Convert a string to the canonical cases. Does not modify the
original string. If you need case-insensitive string
comparisons that follow the equality rules of the server, use
this function to canonize the strings before comparing them.
Check if a particular character is in a string.
Getting and setting global variables
Set the value of a variable in the LysKOM buffer associated
with the current buffer. This is useful if a function in, say,
the edit buffer wants to set a variable in the LysKOM session,
not just in the edit buffer.
Return the value of a variable in the LysKOM buffer associated
with the current buffer. This is useful if a function in, say,
the edit buffer needs the value of a variable that is not
inherited (which in turn is necessary if the edit buffer needs
to know the current value, not just the value at the time the
edit buffer was created).
This is just like lyskom-default-value, but instead of
signalling an error if the variable is unbound in the LysKOM
buffer, this function just returns nil.
Prefix arguments
Use for all functions that want a user-supplied text number as
their argument. Read the docstring.
These functions behave just like their non-lyskom
counterparts, but do their job in the LysKOM buffer associated
with the current buffer. This is useful for code that runs in
one buffer but needs to modify or run hooks in the main LysKOM
Magic wrappers
These don't fit in any particular category, but you need to know
Run the body, but ignore any asynchronous messages that match
a particular pattern while the body is running. This can be
used when you do a server call and know you'll se an async
message that you will handle in a different manner, and would
rather like that it isn't handled twice. Be careful not to
filter out messages that you *do* want. Make the pattern
Like save-excursion but does not save point and mark, just the
current buffer. When you're thinking about doing
save-excursion and then running stuff that might change point
in the LysKOM buffer (such as displaying a text), perhaps this
is what you mean.
This macro surrounds every single server call implementation.
The functions in services.el show how it's used.
Run the body after setting up the minibuffer according to the
LysKOM configuration. Anything you do where the user may
interact with the minibuffer using any non-lyskom functions
(such as read-from-minibuffer or completing-read) should be
surrounded by this.
Text insertion
All these functions insert text in the current buffer. The
"-insert" functions insert the text at the end of the buffer.
The "-insert-at-point" functions insert the text at point. The
"-insert-before-prompt" functions insert the text before the
current prompt (or next prompt for that matter).
The functions with "format" in their names run lyskom-format
on the arguments. (lyskom-format-insert X) is essentially the
same as (lyskom-insert (lyskom-format X)), but the former
allows deferred printing and special insertion (such as HTML)
and is generally faster and nicer than latter piece of code.
Always use lyskom-format-insert if you can.
All functions may cause the buffer to scroll (I think)
according to the inscrutable rules of LysKOM scrolling.
Defining commands
See the file named "commands" for more information.
Defining variables
See the file named "variables" for more information.
Buffer management
See the file named "buffers" for more information.
Text formatting
Se the file named "format" for more information.
David Byers
This file explains how to define global variables in LysKOM so
that they behave the way they are supposed to. It also explains
how to define variables whose values depend on the current
All developers need to know this.
Variables should be named kom-something or lyskom-something. The
kom prefix is for variables that the user is expected to play
with. These should also be available in the customization buffer.
The lyskom prefix is for internal variables.
Define global variables using the def-kom-var macro:
(def-kom-var NAME VALUE &rest PROPERTIES)
NAME is the name of the variable and VALUE is its initial value,
just like in defvar. Unlike defvar, you must provide an initial
value. PROPERTIES are properties of the variable. A string
property is the documentation for the variable, a symbol is one of
the predefined properties listed below and a list is a widget
specification for the variable (see below).
The following predefined properties are available:
server Save the variable in the elisp block of the user's
user-area. All user-configurable variables should have
the server property specified, even if it makes more
senst to save the value locally, since recent versions
of the client allow the user to specify which
variables are stored in the server and which are
stored locally. This property implies the local
local The variable is buffer-local.
inherited The value of the variable is inherited from the parent
buffer when the LysKOM buffer management functions are
used to create a new buffer. The bindings are made
buffer local. This property implies the protected and
local properties.
protected The variable is marked as permanent local, and is not
affected by kill-all-local-variables. This property
implies the local property.
minibuffer The variable is inherited as a local variable in the
minibuffer. Use this for variables that need to be
available to minibuffer hooks and the like.
server-hook The variable is a hook function that is stored in the
local-hook The variable is a hook variable that is to be made
buffer local.
An element of the PROPERTIES argumens to def-kom-var that is a
list is considered a widget specification for the variable.
Currently all widget specifications are listed in option-edit.el
rather than with the variable definitions, so this functionality
is not tested. Feel free to test it or even to move the widget
specifications from option-edit.el to the definitions of the
corresponding variables.
If you need a variable whose value is language-dependent, use the
lyskom-language-var function to set its initial value. The
variable still needs to be defined with def-kom-var, but the
default value will be replaced by the language-specific value with
the language is set.
(lyskom-language-var NAME LANGUAGE VALUE)
NAME is the unquoted name of the variable, as in def-kom-var.
LANGUAGE is the language the value is valid for, also an unquoted
symbol. VALUE is the default value of the variable for the
specified language. The VALUE parameter is evaluated.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment