pam-mode.texi from CQual at Krugle
Show pam-mode.texi syntax highlighted
\input texinfo @c -*-texinfo-*-
@c %**start of header
@setfilename pam-mode.info
@settitle Program Analysis Mode (PAM) version 3.0
@setchapternewpage odd
@c %**end of header
@ifinfo
This info page describes the Program Analysis mode (PAM, for short) version 3.0 for
Emacs.
Copyright @copyright{} 1999-2001 The Regents of the University of California.
PAM is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.
PAM is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with PAM; see the file COPYING. If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.
@end ifinfo
@titlepage
@title Program Analysis Mode
@subtitle Version 3.0
@author Chris Harrelson @email{chrishtr@@cs.berkeley.edu}, Henning Niss @email{hniss@@cs.berkeley.edu}
@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1999-2001 The Regents of the University of California.
@end titlepage
@comment node-name, next, prev, up
@node top, Copying, , (dir)
@ifinfo
@top Program Analysis Mode
This info page describes the Program Analysis Mode (PAM, for short), a
minor mode for Emacs.
It applies to version 3.0 of both the Emacs Lisp and ML portions of code contained
in the package.
@end ifinfo
@menu
* Copying:: Your rights
* Installation:: What to do before you start
* Start:: What to do first
* Source:: How to set things up for SML/NJ and other compilers
* Variables:: Variable definitions
* Functions:: Functions
* Keybindings:: Keyboard and mouse bindings
* Contact:: How to contact us
@end menu
@comment node-name, next, previous, up
@node Copying, Installation, top, top
@quotation
Copyright @copyright{} 1999-2001 The Regents of the University of California.
PAM is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.
PAM is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with PAM; see the file COPYING. If not, write to
the Free Software Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.
@end quotation
@comment node-name, next, previous, up
@node Installation, Start, Copying, top
@chapter Installation
For information about installation, see the file
'README', or run 'make' in the PAM source directory.
@c -----------------------
@node Start, Source, Installation, top
@chapter Getting started
The first thing you need to do is set up your @file{.emacs} file for
the analysis that you would like to run. Each analysis should provide
appropriate commands to set it up in its own documentation. However,
here is an example in case you want to make your own.
This analysis is called @samp{simple} and should have come with your PAM
distribution in a sub-directory. Before continuing, you should have
installed the SML/NJ compiler and run the command @file{make simple} in
the directory containing the PAM distribution. This produces an
executable binary called @file{simple.ARCH}, where @file{ARCH} depends on the
kind of computer you are running. For example, on a Linux machine, this
is usually @file{x86-linux}.
Next, add these lines to your @file{.emacs} file (warning: this assumes you have
already installed PAM correctly according to the instructions in the
@ref{Installation} section):
@example
; [1]
(setq pam-default-analysis '("sml-cm" "@@SMLload=SIMPLE_PATH/simple.ARCH"
"SIMPLE_PATH/dictionary"))
; [2]
(setq pam-default-sml-analysis '("SIMPLE_PATH/simple"
"SIMPLE_PATH/dictionary"))
@end example
In the Emacs Lisp lines above, you need to substitute the full path
where the @file{simple} example analysis binary is stored wherever it
says @file{SIMPLE_PATH}. Remember that you should also replace
@file{ARCH} with the extension SML/NJ gives for your particular type of
computer.
Now you need to load these changes into Emacs. This can be accomplished
by restarting Emacs. After the changes have been read by Emacs, you
should be able to type three different commands to get the analysis
started:
@example
[3] @samp{M-x pam-analyze-file}
[4] @samp{M-x pam-sml-analyze-file}
[5] @samp{M-x pam-full-analyze-file}
@end example
Typing [3] (i.e. @samp{M-x pam-analyze-file}, not the "[3] " part) will
use the information you gave above in your @file{.emacs} file to
initialize the system. It will then arrive at a prompt for more
information saying "File to analyze: ". Enter the file to be analyzed
(for the @samp{simple} analysis, any text file will do). If all goes
well, the Table of Contents screen will appear with a summary analysis
of the input file you provided. Items there that are underlined should
be clickable (with the middle mouse button, or if you prefer keyboard
interaction you can move the cursor on top of a highlighted word and
type @samp{C-c C-l}).
On the other hand, if you had executed [4] or [5], you would have been
prompted one extra time at the beginning @strong{before} the "File to
analyze: " prompt. They ask for more specific information if you need
to give it: @code{pam-sml-analyze-file} expects its first input to be something
with the same syntax as the @code{pam-default-sml-analysis} variable you set
above (see the comment [2]), and @code{pam-full-analyze-file expects} (you
guessed it) something with the same syntax as the @code{pam-default-analysis}
variable. See @ref{Variables} for more details. If possible, use pam-sml-analyze-file,
since its inputs are simpler.
That's pretty much it as far as the Emacs and user side goes. Various
configuration variables and key bindings, along with the semantics of
the PAM "minor mode", are described in the next sections, but you can
probably guess at them with a little exploration.
@c -----------------
@node Source, Variables, Start, top
@chapter Analysis source types
The only SML compiler that we have tested with PAM is
SML/NJ. Other compilers should work with minor changes.
We assume several things for @code{pam-sml-analyze-file}
to work correctly:
@enumerate
@item The sml program (given by the global variable @code{sml-program-name};
see @ref{Variables}) is in the operating system path.
@item Your system configuration is one of the ones we support directly,
or you have put yours into the configuration table variable (again,
see @ref{Variables}).
@item Your version of SML/NJ is compatible with the SML'97 way of doing
things. It must support:
@itemize
@item The "@@SMLload" argument to load a binary
@item The use of specific system configuration extensions on binary
names. For instance, on an x86 Linux system, the @samp{simple}
analysis' binary is called "simple.x86-linux"
@end itemize
@end enumerate
If any of 1, 2 or 3 are not the case, you will probably have to use
pam-full-analyze-file, since it makes no assumptions about your
system at all.
We have not yet tested the system with any other SML compilers (MLWorks,
MLTon, MLj, MLKit, etc.). However, it should be trivial to compile
binaries and connect them to the Emacs interface with the appropriate
setting of @code{pam-full-analyze-file}. The only things really
necessary are the ability of Emacs to connect to the binary as a child
process and have their stdout/stdin interfaces thereby connected. If
you come up with the right way to do it, please send us your
configuration so it can be added to the interface in the next version.
@c -------------------------
@node Variables, Functions, Source, top
@chapter PAM system variables
@itemize
@item @code{pam-default-analysis} (default: nil)
This is the default analysis used by both
@code{pam-full-analyze-file} and @code{pam-analyze-file}. It has the
format:
@example
"program" "arg1" "arg2" ... "argn"
@end example
It is very important that all members of this list are proper strings.
@code{program} = The name of the program to run; if it isn't in the operating
system path you will need to specify an absolute one.
@code{argN} (for all N) = The text of the Nth argument to pass to program.
Note that it will be globbed by the file system.
@item @code{pam-sml-default-analysis} (default: nil)
This is the default initialization used by @code{pam-sml-analyze-file}. It
has the format:
@example
("binary-file-name-prefix" "arg1" "arg2" ... "argn")
@end example
@code{binary-file-name-prefix} = Everything in the name of the SML/NJ binary
file preceding the dot (".") and the architecture-specific extension.
For example, in the @samp{simple} analysis, the complete file name might be
"simple.x86-linux" while the prefix is just "simple". The advantage of not
specifying this precisely is that if you have a heterogeneous computing
environment (like we do: our lab has Sun, HP, DEC and Intel systems) no
special configuration (other than recompiled binaries) will be needed on
specific machines.
@item @code{pam-results-buf} (default "*pam-results-buf*")
The name of the buffer where raw data from the analysis program appears.
Rename this if you like.
@item @code{pam-debug-buf} (default: "*pam-debug-buf*")
The name of the buffer where debugging information from the analysis
program appears. Rename this if you like.
@item @code{sml-program-name} (default: "sml-cm")
Used as a shortcut when running sml analyses. The version bound by the
sml-mode package will automatically override this if it's loaded first.
@item @code{pam-sml-system-configuration-table} (default: nil)
A list of cons cells, each of which has a regular expression against
which to match in the car field, and the extension to put after a file
name in the cdr field. This is used to deduce the filename extension to
be used in @code{pam-sml-analyze-file}. The regular expressions match
against the global Emacs variable @code{system-configuration}. Only
insert values in this table if the predefined ones don't work for you.
This list is checked @strong{before} any of the built-in heuristics are
performed.
@item @code{pam-color-n}, where n is in the set [1, ..., 8]
The eight different PAM colors that can be chosen by the SML process
for display. You may need to configure these if the standard colors
look terrible. They can be changed by simply setting them, as in:
@example
(setq pam-color-1 'highlight)
@end example
This will make pam-color-1 have the same color as the built-in
Emacs face @samp{highlight}. Alternatively, you can use the
@code{custom-set-faces} function to create your own faces (see
the Emacs documentation for more details). (Note: default
values intentionally left blank)
@item @code{pam-highlight-color} (default: 'highlight)
The face PAM uses to draw a highlighted element in a PAM file, as
directed by the SML process. You can set this to be any color you like
(see just above).
@end itemize
@c ------------------------------
@node Functions, Keybindings, Variables, top
@chapter Functions
Notes on interactive entries in the Emacs minibuffer:
@itemize
@item If a file name is expected (for example, with the last
prompt of "File to analyze: "), filename completion is provided for you,
through the use of [TAB] as usual. However, the final entered string
_does not_ have to be an actual file name or directory. You can
also use the history list described below.
@item If it is not a file name (for example, with the first prompt of
@code{pam-sml-analyze-file} or @code{pam-full-analyze-file}), you can use
a history list kept specifically for those inputs. Use @samp{M-p} to go
up the list and @samp{M-n} to go down. Consult the Emacs documentation
for more details, since this is a standard feature of Emacs input buffers.
You can execute these functions in the usual Emacs way by typing
@samp{M-x function-name}:
@item @code{pam-analyze-file}
This uses the @code{pam-default-analysis} global variable to decide what
analysis to use, then prompts for the name of the data to pass to it.
@item @code{pam-sml-analyze-file}
This is the same as @code{pam-analyze-file}, except that it asks you
first for a list structure to execute (see the description of
@code{pam-sml-default-analysis} in @ref{Variables}, for details on the
input format).
@item @code{pam-full-analyze-file}
This is the same as @code{pam-sml-analyze-file}, except that it asks you
for a list in the format of pam-default-analysis instead.
@item @code{pam-mode}
Toggles whether the PAM minor mode is active in the current buffer.
@item @code{pam-reset}
Resets the PAM system, including emptying internal variables and closing
all PAM-opened buffers. Note that if you have a source buffer open
before PAM is run, it will remain open after @code{pam-reset} has been
executed, while if PAM opens the buffer itself, it will be closed. This
function is automatically called by the @samp{pam-*-analyze-file}
functions.
@end itemize
@c -------------------------
@node Keybindings, Contact, Functions, top
@chapter Keybindings and mouse bindings
Note that these key and mouse bindings will only work in a buffer with
the PAM minor mode running. You can tell if it is running by checking
if it says "PAM" in the status line of the current buffer.
Keyboard bindings:
@itemize
@item @samp{C-c C-f}
A shorthand for @code{pam-analyze-file}.
@item @samp{C-c C-r}
A shorthand for @code{pam-reset}.
@item @samp{C-c C-l}
Does the equivalent of @emph{clicking} on the text under the cursor.
If that text is a hyperlink, it will be followed.
@item @samp{C-c C-i}
Loads the GNU info documentation for PAM.
@item @samp{C-c C-m}
Toggles PAM mode in the current buffer.
@end itemize
Mouse Bindings:
@itemize
@item Middle mouse button
Activates a "click" on the character underneath the mouse cursor. If
the character is part of a clickable hyperlink, then that hyperlink will
be followed.
@item Right mouse button
Activates a "click" on the region currently highlighted in the buffer.
For example, if you highlight a specific word, then click the right
mouse button, PAM will look for the hyperlink that most closely matches
the highlighted region, and execute that hyperlink. This functionality
is useful when hyperlinks overlap in the file, and hence simply
clicking on a hyperlink becomes ambiguous.
@end itemize
@c ---------------------
@node Contact, , Keybindings, top
@chapter How to contact us
The authors, by email:
Chris Harrelson @email{chrishtr@@cs.berkeley.edu}
Henning Niss
The web page:
@url{http://www.cs.berkeley.edu/~chrishtr/pam}
@bye
@c LocalWords: texinfo setfilename pam settitle setchapternewpage ifinfo Niss
@c LocalWords: MERCHANTABILITY HEREUNDER titlepage Henning vskip filll prev
@c LocalWords: dir SML Keybindings README emacs samp linux setq sml clickable
@c LocalWords: pre MLWorks MLTon MLj MLKit stdout stdin arg argn argN globbed
@c LocalWords: buf cdr filename minibuffer emph hyperlink hyperlinks url
See more files for this project here