*** empty log message ***

[gnu-emacs] / lispref / anti.texi

@c -*-texinfo-*-
@c This is part of the GNU Emacs Lisp Reference Manual.
@c Copyright (C) 1998 Free Software Foundation, Inc. 
@c See the file elisp.texi for copying conditions.
@node Antinews, Index, Standard Hooks, Top
@appendix Emacs 19 Antinews

For those users who live backwards in time, here is information about
downgrading to Emacs version 19.34.  We hope you will enjoy the greater
simplicity that results from the absence of many Emacs 19 features.  In
the following section, we carry this information back as far as Emacs
19.29, for which the previous printed edition of this manual was made.

@section Old Lisp Features in Emacs 19

Here are the most important of the features that you will learn
to do without in Emacs 19:

@itemize @bullet
@item
In a great simplification, Emacs 19 supports ASCII characters only.
There are no multibyte characters, character sets, language
environments, coding systems, or input methods; all the functions that
specifically relate to them are gone as well.

Valid character codes for text must be in the range 0 through 255.
Within this range, there are no invalid character codes.

@item
The Custom facility has been replaced with a much simpler and more
general method of defining user option variables.  Instead of
@code{defcustom}, which requires you to specify each user option's
data type and classify them into groups, all you have to do is write
a @code{defvar} and start the documentation string with @samp{*}.
@end itemize

Here are changes in the Lisp language itself:

@itemize @bullet
@item
Symbols whose names start with @samp{:} are no longer special
in any way.  They start out void, like most other symbols.

@item
The macros @code{when} and @code{unless} have been deleted.

@item
The functions @code{caar}, @code{cadr}, @code{cdar} and @code{cddr}
no longer exist.

@item
The function @code{functionp} is now gone.  If you don't know
by now whether something is a function, Emacs can't tell you.
@end itemize

Here are changes in handling strings and text.

@itemize @bullet
@item
The function @code{substring} works only on strings, not on vectors.

@item
There are no more character categories.

@item
When you compare strings with @code{equal}, it now compares
their string properties as well as their text.  All must match,
or the strings are not equal.

@item
@code{format-time-string} no longer supports specified field width
or specified padding.

@item
The functions @code{split-string} and @code{concat-chars} no longer exist.
Neither does @code{store-substring} or @code{sref}.

@item
All printing characters have the same width.  Therefore, we have deleted
@code{char-width}, @code{string-width} and
@code{truncate-string-to-width}.

@item
We have eliminated the functions @code{next-char-property-change} and
@code{previous-char-property-change} also.

@item
Syntax parsing now determines the syntax of each character from the
syntax table alone---not from text properties.  This makes the syntax
codes @samp{|} and @samp{!}, which were meant for use with text
properties, useless; so we have deleted them.

@item
In the function @code{parse-partial-sexp}, passing @code{syntax-table}
as the sixth argument @var{commentstop} no longer has any special meaning.
And the return value has only eight elements.
@end itemize

Here are changes in other areas of Emacs Lisp:

@itemize @bullet
@item
The macros @code{save-current-buffer}, @code{with-current-buffer},
@code{with-temp-buffer}, @code{with-temp-file}, @code{save-selected-window}, 
and @code{with-output-to-string} are gone.

@item
The easy-mmode facility for defining minor modes is gone too.

@item
Process filters and sentinels must explicitly save the match data, with
@code{save-match-data}, or they will clobber the match data and
something horrible will happen.

@item
As part of our effort to loosen up, @code{batch-byte-compile-file} no
longer returns a nonzero status code if there is a compilation error.

@item
The ``mail user agent'' feature is gone.

@item
We have removed the functions @code{add-to-invisibility-spec} and
@code{remove-from-invisibility-spec}, so you should manipulate
the value of @code{buffer-invisibility-spec} by hand.

@item
The functions @code{face-documentation}, @code{face-bold-p},
@code{face-italic-p}, @code{set-face-bold-p}, @code{set-face-italic-p}
are gone.  Instead, use @code{make-face-bold} and friends.

@item
All the functions that operate on a file now discard an extra redundant
directory name from the beginning of the file name---just like
@code{substitute-in-file-name}.

@item
We have got rid of the function @code{access-file}.

@item
Most of the minibuffer input functions, no longer take a default value as
an argument.  Also, they do not discard text properties from the result.
This means that if you insert text with text properties into the minibuffer,
the minibuffer value really will contain text properties.

@item
You can still bind @code{x-resource-class} around a call to
@code{x-get-resource}, but it won't do anything special.

@item
Wave goodbye to the hooks @code{before-make-frame-hook},
@code{after-make-frame-functions}, and
@code{window-configuration-change-hook},

@item
The functions and variables that deal with MS Windows NT/95
have been renamed to start with @samp{win32-} instead of @samp{w32-}.
This is because we admire Microsoft more each day as we go back
into the past.
@end itemize

@section Onward into the Past!

Here we go even further back, as far as Emacs 19.29, for which the
previous printed edition of the Emacs Lisp manual was made.

@itemize @bullet
@item
There are no char-tables or bool-vectors.  Syntax tables, display
tables, and case tables are all vectors now, and the value of
@code{keyboard-translate-table} should be a vector or a string.

@item
There is only one kind of marker.  When you insert text at the place
where a marker points, the marker always ends up before that text,
unless you use @code{insert-before-markers}, which puts all the markers
after the inserted text.

@item
There is no function @code{overlays-in}.

@item
The variable @code{print-length} applies only to lists, not to
vectors or strings.

@item
The function @code{convert-standard-filename} no longer exists, so each
Lisp package must independently figure out which file names to use for
its initialization files on each kind of operating system.

@item
The macro @code{with-timeout} has been eliminated, along with the
function @code{y-or-n-p-with-timeout}.  Idle timers don't exist at all;
instead, maybe you can use @code{post-command-idle-hook} to do some of
the same job.

@item
The functions @code{keymap-parent} and @code{set-keymap-parent} are
gone.  We expect keymaps to recognize their own parents.

@item
When you delete text and then undo a deletion, markers that were
originally inside the deleted text end up either at the beginning
or the end of it---not back in their original places.

@item
The interactive specification @samp{N} is gone now.

@item
There is no more @code{safe-length}.  Don't try to be so safe!  Did you
expect to live forever?

@item
We got rid of @code{insert-file-contents-literally}, because
programmers are too literal-minded anyway.

@item
As part of our continuing effort to help Lisp programmers to relax, we
threw out the function @code{error-message-string}.  Don't worry so much
about errors!  We all make mistakes.

@item
The keymap @code{special-event-map} is gone, because Emacs has no more
special events.  If you want to hold a party in Emacs, please let us
know.

@item
You can't do date arithmentic with @code{encode-time} any more.

@item
The functions @code{command-execute} and @code{call-interactively} no
longer accept the optional argument @var{keys}.

@item
@code{get-buffer-window-list} is gone as well.

@item
With the function @code{replace-match}, you can only replace the whole
match, not a subexpression of it.

@item
We eliminated the hooks @code{buffer-access-fontify-functions},
@code{window-scroll-functions}, and @code{redisplay-end-trigger-functions}.
@end itemize