Blame - docs/Stacker.html - platform/external/llvm

blob: 81ad60e8fd609bbbc55b46ce4053f5a11b82dcf3 [file] [log] [blame]

Brian Gaeke	9018148	2003-11-24 02:52:51 +0000	[diff] [blame]	1	<!DOCTYPE HTML PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
				2	<html>
				3	<head>
				4	<title>Stacker: An Example Of Using LLVM</title>
				5	<link rel="stylesheet" href="llvm.css" type="text/css">
				6	</head>
				7	<body>
				8	<div class="doc_title">Stacker: An Example Of Using LLVM</div>
				9	<ol>
				10	<li><a href="#abstract">Abstract</a></li>
				11	<li><a href="#introduction">Introduction</a></li>
				12	<li><a href="#lexicon">The Stacker Lexicon</a>
				13	<ol>
				14	<li><a href="#stack">The Stack</a>
				15	<li><a href="#punctuation">Punctuation</a>
				16	<li><a href="#literals">Literals</a>
				17	<li><a href="#words">Words</a>
				18	<li><a href="#builtins">Built-Ins</a>
				19	</ol>
				20	</li>
				21	<li><a href="#directory">The Directory Structure </a>
				22	</ol>
				23	<div class="doc_text">
				24	<p><b>Written by <a href="mailto:rspencer@x10sys.com">Reid Spencer</a> </b></p>
				25	<p> </p>
				26	</div>
				27	<!-- ======================================================================= -->
				28	<div class="doc_section"> <a name="abstract">Abstract </a></div>
				29	<div class="doc_text">
				30	<p>This document is another way to learn about LLVM. Unlike the
				31	<a href="LangRef.html">LLVM Reference Manual</a> or
				32	<a href="ProgrammersManual.html">LLVM Programmer's Manual</a>, this
				33	document walks you through the implementation of a programming language
				34	named Stacker. Stacker was invented specifically as a demonstration of
				35	LLVM. The emphasis in this document is not on describing the
				36	intricacies of LLVM itself, but on how to use it to build your own
				37	compiler system.</p>
				38	</div>
				39	<!-- ======================================================================= -->
				40	<div class="doc_section"> <a name="introduction">Introduction</a> </div>
				41	<div class="doc_text">
				42	<p>Amongst other things, LLVM is a platform for compiler writers.
				43	Because of its exceptionally clean and small IR (intermediate
				44	representation), compiler writing with LLVM is much easier than with
				45	other system. As proof, the author of Stacker wrote the entire
				46	compiler (language definition, lexer, parser, code generator, etc.) in
				47	about <em>four days</em>! That's important to know because it shows
				48	how quickly you can get a new
				49	language up when using LLVM. Furthermore, this was the <em >first</em>
				50	language the author ever created using LLVM. The learning curve is
				51	included in that four days.</p>
				52	<p>The language described here, Stacker, is Forth-like. Programs
				53	are simple collections of word definitions and the only thing definitions
				54	can do is manipulate a stack or generate I/O. Stacker is not a "real"
				55	programming language; its very simple. Although it is computationally
				56	complete, you wouldn't use it for your next big project. However,
				57	the fact that it is complete, its simple, and it <em>doesn't</em> have
				58	a C-like syntax make it useful for demonstration purposes. It shows
				59	that LLVM could be applied to a wide variety of language syntaxes.</p>
				60	<p>The basic notions behind stacker is very simple. There's a stack of
				61	integers (or character pointers) that the program manipulates. Pretty
				62	much the only thing the program can do is manipulate the stack and do
				63	some limited I/O operations. The language provides you with several
				64	built-in words that manipulate the stack in interesting ways. To get
				65	your feet wet, here's how you write the traditional "Hello, World"
				66	program in Stacker:</p>
				67	<p><code>: hello_world "Hello, World!" >s DROP CR ;<br>
				68	: MAIN hello_world ;<br></code></p>
				69	<p>This has two "definitions" (Stacker manipulates words, not
				70	functions and words have definitions): <code>MAIN</code> and <code>
				71	hello_world</code>. The <code>MAIN</code> definition is standard, it
				72	tells Stacker where to start. Here, <code>MAIN</code> is defined to
				73	simply invoke the word <code>hello_world</code>. The
				74	<code>hello_world</code> definition tells stacker to push the
				75	<code>"Hello, World!"</code> string onto the stack, print it out
				76	(<code>>s</code>), pop it off the stack (<code>DROP</code>), and
				77	finally print a carriage return (<code>CR</code>). Although
				78	<code>hello_world</code> uses the stack, its net effect is null. Well
				79	written Stacker definitions have that characteristic. </p>
				80	<p>Exercise for the reader: how could you make this a one line program?</p>
				81	</div>
				82	<!-- ======================================================================= -->
				83	<div class="doc_section"><a name="stack"></a>Lessons Learned About LLVM</div>
				84	<div class="doc_text">
				85	<p>Stacker was written for two purposes: (a) to get the author over the
				86	learning curve and (b) to provide a simple example of how to write a compiler
				87	using LLVM. During the development of Stacker, many lessons about LLVM were
				88	learned. Those lessons are described in the following subsections.<p>
				89	</div>
				90	<div class="doc_subsection"><a name="linkage"></a>Getting Linkage Types Right</div>
				91	<div class="doc_text"><p>To be completed.</p></div>
				92	<div class="doc_subsection"><a name="linkage"></a>Everything's a Value!</div>
				93	<div class="doc_text"><p>To be completed.</p></div>
				94	<div class="doc_subsection"><a name="linkage"></a>The Wily GetElementPtrInst</div>
				95	<div class="doc_text"><p>To be completed.</p></div>
				96	<div class="doc_subsection"><a name="linkage"></a>Constants Are Easier Than That!</div>
				97	<div class="doc_text"><p>To be completed.</p></div>
				98	<div class="doc_subsection"><a name="linkage"></a>Terminate Those Blocks!</div>
				99	<div class="doc_text"><p>To be completed.</p></div>
				100	<div class="doc_subsection"><a name="linkage"></a>new,get,create .. Its All The Same</div>
				101	<div class="doc_text"><p>To be completed.</p></div>
				102	<div class="doc_subsection"><a name="linkage"></a>Utility Functions To The Rescue</div>
				103	<div class="doc_text"><p>To be completed.</p></div>
				104	<div class="doc_subsection"><a name="linkage"></a>push_back Is Your Friend</div>
				105	<div class="doc_text"><p>To be completed.</p></div>
				106	<div class="doc_subsection"><a name="linkage"></a>Block Heads Come First</div>
				107	<div class="doc_text"><p>To be completed.</p></div>
				108	<!-- ======================================================================= -->
				109	<div class="doc_section"> <a name="lexicon">The Stacker Lexicon</a></div>
				110	<div class="doc_subsection"><a name="stack"></a>The Stack</div>
				111	<div class="doc_text">
				112	<p>Stacker definitions define what they do to the global stack. Before
				113	proceeding, a few words about the stack are in order. The stack is simply
				114	a global array of 32-bit integers or pointers. A global index keeps track
				115	of the location of the to of the stack. All of this is hidden from the
				116	programmer but it needs to be noted because it is the foundation of the
				117	conceptual programming model for Stacker. When you write a definition,
				118	you are, essentially, saying how you want that definition to manipulate
				119	the global stack.</p>
				120	<p>Manipulating the stack can be quite hazardous. There is no distinction
				121	given and no checking for the various types of values that can be placed
				122	on the stack. Automatic coercion between types is performed. In many
				123	cases this is useful. For example, a boolean value placed on the stack
				124	can be interpreted as an integer with good results. However, using a
				125	word that interprets that boolean value as a pointer to a string to
				126	print out will almost always yield a crash. Stacker simply leaves it
				127	to the programmer to get it right without any interference or hindering
				128	on interpretation of the stack values. You've been warned :) </p>
				129	</div>
				130	<!-- ======================================================================= -->
				131	<div class="doc_subsection"> <a name="punctuation"></a>Punctuation</div>
				132	<div class="doc_text">
				133	<p>Punctuation in Stacker is very simple. The colon and semi-colon
				134	characters are used to introduce and terminate a definition
				135	(respectively). Except for <em>FORWARD</em> declarations, definitions
				136	are all you can specify in Stacker. Definitions are read left to right.
				137	Immediately after the semi-colon comes the name of the word being defined.
				138	The remaining words in the definition specify what the word does.</p>
				139	</div>
				140	<!-- ======================================================================= -->
				141	<div class="doc_subsection"><a name="literals"></a>Literals</div>
				142	<div class="doc_text">
				143	<p>There are three kinds of literal values in Stacker. Integer, Strings,
				144	and Booleans. In each case, the stack operation is to simply push the
				145	value onto the stack. So, for example:<br/>
				146	<code> 42 " is the answer." TRUE </code><br/>
				147	will push three values onto the stack: the integer 42, the
				148	string " is the answer." and the boolean TRUE.</p>
				149	</div>
				150	<!-- ======================================================================= -->
				151	<div class="doc_subsection"><a name="words"></a>Words</div>
				152	<div class="doc_text">
				153	<p>Each definition in Stacker is composed of a set of words. Words are
				154	read and executed in order from left to right. There is very little
				155	checking in Stacker to make sure you're doing the right thing with
				156	the stack. It is assumed that the programmer knows how the stack
				157	transformation he applies will affect the program.</p>
				158	<p>Words in a definition come in two flavors: built-in and programmer
				159	defined. Simply mentioning the name of a previously defined or declared
				160	programmer-defined word causes that words definition to be invoked. It
				161	is somewhat like a function call in other languages. The built-in
				162	words have various effects, described below.</p>
				163	<p>Sometimes you need to call a word before it is defined. For this, you can
				164	use the <code>FORWARD</code> declaration. It looks like this</p>
				165	<p><code>FORWARD name ;</code></p>
				166	<p>This simply states to Stacker that "name" is the name of a definition
				167	that is defined elsewhere. Generally it means the definition can be found
				168	"forward" in the file. But, it doesn't have to be in the current compilation
				169	unit. Anything declared with <code>FORWARD</code> is an external symbol for
				170	linking.</p>
				171	</div>
				172	<!-- ======================================================================= -->
				173	<div class="doc_subsection"><a name="builtins"></a>Built In Words</div>
				174	<div class="doc_text">
				175	<p>The built-in words of the Stacker language are put in several groups
				176	depending on what they do. The groups are as follows:</p>
				177	<ol>
				178	<li><em>Logical</em>These words provide the logical operations for
				179	comparing stack operands.<br/>The words are: < > <= >=
				180	= <> true false.</li>
				181	<li><em>Bitwise</em>These words perform bitwise computations on
				182	their operands. <br/> The words are: << >> XOR AND NOT</li>
				183	<li><em>Arithmetic</em>These words perform arithmetic computations on
				184	their operands. <br/> The words are: ABS NEG + - * / MOD */ ++ -- MIN MAX</li>
				185	<li><em>Stack</em>These words manipulate the stack directly by moving
				186	its elements around.<br/> The words are: DROP DUP SWAP OVER ROT DUP2 DROP2 PICK TUCK</li>
				187	<li><em>Memory></em>These words allocate, free and manipulate memory
				188	areas outside the stack.<br/>The words are: MALLOC FREE GET PUT</li>
				189	<li><em>Control</em>These words alter the normal left to right flow
				190	of execution.<br/>The words are: IF ELSE ENDIF WHILE END RETURN EXIT RECURSE</li>
				191	<li><em>I/O</em> These words perform output on the standard output
				192	and input on the standard input. No other I/O is possible in Stacker.
				193	<br/>The words are: SPACE TAB CR >s >d >c <s <d <c.</li>
				194	</ol>
				195	<p>While you may be familiar with many of these operations from other
				196	programming languages, a careful review of their semantics is important
				197	for correct programming in Stacker. Of most importance is the effect
				198	that each of these built-in words has on the global stack. The effect is
				199	not always intuitive. To better describe the effects, we'll borrow from Forth the idiom of
				200	describing the effect on the stack with:</p>
				201	<p><code> BEFORE -- AFTER </code></p>
				202	<p>That is, to the left of the -- is a representation of the stack before
				203	the operation. To the right of the -- is a representation of the stack
				204	after the operation. In the table below that describes the operation of
				205	each of the built in words, we will denote the elements of the stack
				206	using the following construction:</p>
				207	<ol>
				208	<li><em>b</em> - a boolean truth value</li>
				209	<li><em>w</em> - a normal integer valued word.</li>
				210	<li><em>s</em> - a pointer to a string value</li>
				211	<li><em>p</em> - a pointer to a malloc's memory block</li>
				212	</ol>
				213	</div>
				214	<div class="doc_text">
				215	<table class="doc_table" >
				216	<tr class="doc_table"><td colspan="4">Definition Of Operation Of Built In Words</td></tr>
				217	<tr class="doc_table"><td colspan="4">LOGICAL OPERATIONS</td></tr>
				218	<tr class="doc_table"><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				219	<tr class="doc_table"><td><</td>
				220	<td>LT</td>
				221	<td>w1 w2 -- b</td>
				222	<td>Two values (w1 and w2) are popped off the stack and
				223	compared. If w1 is less than w2, TRUE is pushed back on
				224	the stack, otherwise FALSE is pushed back on the stack.</td>
				225	</tr>
				226	<tr><td>></td>
				227	<td>GT</td>
				228	<td>w1 w2 -- b</td>
				229	<td>Two values (w1 and w2) are popped off the stack and
				230	compared. If w1 is greater than w2, TRUE is pushed back on
				231	the stack, otherwise FALSE is pushed back on the stack.</td>
				232	</tr>
				233	<tr><td>>=</td>
				234	<td>GE</td>
				235	<td>w1 w2 -- b</td>
				236	<td>Two values (w1 and w2) are popped off the stack and
				237	compared. If w1 is greater than or equal to w2, TRUE is
				238	pushed back on the stack, otherwise FALSE is pushed back
				239	on the stack.</td>
				240	</tr>
				241	<tr><td><=</td>
				242	<td>LE</td>
				243	<td>w1 w2 -- b</td>
				244	<td>Two values (w1 and w2) are popped off the stack and
				245	compared. If w1 is less than or equal to w2, TRUE is
				246	pushed back on the stack, otherwise FALSE is pushed back
				247	on the stack.</td>
				248	</tr>
				249	<tr><td>=</td>
				250	<td>EQ</td>
				251	<td>w1 w2 -- b</td>
				252	<td>Two values (w1 and w2) are popped off the stack and
				253	compared. If w1 is equal to w2, TRUE is
				254	pushed back on the stack, otherwise FALSE is pushed back
				255	</td>
				256	</tr>
				257	<tr><td><></td>
				258	<td>NE</td>
				259	<td>w1 w2 -- b</td>
				260	<td>Two values (w1 and w2) are popped off the stack and
				261	compared. If w1 is equal to w2, TRUE is
				262	pushed back on the stack, otherwise FALSE is pushed back
				263	</td>
				264	</tr>
				265	<tr><td>FALSE</td>
				266	<td>FALSE</td>
				267	<td> -- b</td>
				268	<td>The boolean value FALSE (0) is pushed onto the stack.</td>
				269	</tr>
				270	<tr><td>TRUE</td>
				271	<td>TRUE</td>
				272	<td> -- b</td>
				273	<td>The boolean value TRUE (-1) is pushed onto the stack.</td>
				274	</tr>
				275	<tr><td colspan="4">BITWISE OPERATIONS</td></tr>
				276	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				277	<tr><td><<</td>
				278	<td>SHL</td>
				279	<td>w1 w2 -- w1<<w2</td>
				280	<td>Two values (w1 and w2) are popped off the stack. The w2
				281	operand is shifted left by the number of bits given by the
				282	w1 operand. The result is pushed back to the stack.</td>
				283	</tr>
				284	<tr><td>>></td>
				285	<td>SHR</td>
				286	<td>w1 w2 -- w1>>w2</td>
				287	<td>Two values (w1 and w2) are popped off the stack. The w2
				288	operand is shifted right by the number of bits given by the
				289	w1 operand. The result is pushed back to the stack.</td>
				290	</tr>
				291	<tr><td>OR</td>
				292	<td>OR</td>
				293	<td>w1 w2 -- w2\|w1</td>
				294	<td>Two values (w1 and w2) are popped off the stack. The values
				295	are bitwise OR'd together and pushed back on the stack. This is
				296	not a logical OR. The sequence 1 2 OR yields 3 not 1.</td>
				297	</tr>
				298	<tr><td>AND</td>
				299	<td>AND</td>
				300	<td>w1 w2 -- w2&w1</td>
				301	<td>Two values (w1 and w2) are popped off the stack. The values
				302	are bitwise AND'd together and pushed back on the stack. This is
				303	not a logical AND. The sequence 1 2 AND yields 0 not 1.</td>
				304	</tr>
				305	<tr><td>XOR</td>
				306	<td>XOR</td>
				307	<td>w1 w2 -- w2^w1</td>
				308	<td>Two values (w1 and w2) are popped off the stack. The values
				309	are bitwise exclusive OR'd together and pushed back on the stack.
				310	For example, The sequence 1 3 XOR yields 2.</td>
				311	</tr>
				312	<tr><td colspan="4">ARITHMETIC OPERATIONS</td></tr>
				313	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				314	<tr><td>ABS</td>
				315	<td>ABS</td>
				316	<td>w -- \|w\|</td>
				317	<td>One value s popped off the stack; its absolute value is computed
				318	and then pushed onto the stack. If w1 is -1 then w2 is 1. If w1 is
				319	1 then w2 is also 1.</td>
				320	</tr>
				321	<tr><td>NEG</td>
				322	<td>NEG</td>
				323	<td>w -- -w</td>
				324	<td>One value is popped off the stack which is negated and then
				325	pushed back onto the stack. If w1 is -1 then w2 is 1. If w1 is
				326	1 then w2 is -1.</td>
				327	</tr>
				328	<tr><td> + </td>
				329	<td>ADD</td>
				330	<td>w1 w2 -- w2+w1</td>
				331	<td>Two values are popped off the stack. Their sum is pushed back
				332	onto the stack</td>
				333	</tr>
				334	<tr><td> - </td>
				335	<td>SUB</td>
				336	<td>w1 w2 -- w2-w1</td>
				337	<td>Two values are popped off the stack. Their difference is pushed back
				338	onto the stack</td>
				339	</tr>
				340	<tr><td> * </td>
				341	<td>MUL</td>
				342	<td>w1 w2 -- w2*w1</td>
				343	<td>Two values are popped off the stack. Their product is pushed back
				344	onto the stack</td>
				345	</tr>
				346	<tr><td> / </td>
				347	<td>DIV</td>
				348	<td>w1 w2 -- w2/w1</td>
				349	<td>Two values are popped off the stack. Their quotient is pushed back
				350	onto the stack</td>
				351	</tr>
				352	<tr><td>MOD</td>
				353	<td>MOD</td>
				354	<td>w1 w2 -- w2%w1</td>
				355	<td>Two values are popped off the stack. Their remainder after division
				356	of w1 by w2 is pushed back onto the stack</td>
				357	</tr>
				358	<tr><td> */ </td>
				359	<td>STAR_SLAH</td>
				360	<td>w1 w2 w3 -- (w3*w2)/w1</td>
				361	<td>Three values are popped off the stack. The product of w1 and w2 is
				362	divided by w3. The result is pushed back onto the stack.</td>
				363	</tr>
				364	<tr><td> ++ </td>
				365	<td>INCR</td>
				366	<td>w -- w+1</td>
				367	<td>One value is popped off the stack. It is incremented by one and then
				368	pushed back onto the stack.</td>
				369	</tr>
				370	<tr><td> -- </td>
				371	<td>DECR</td>
				372	<td>w -- w-1</td>
				373	<td>One value is popped off the stack. It is decremented by one and then
				374	pushed back onto the stack.</td>
				375	</tr>
				376	<tr><td>MIN</td>
				377	<td>MIN</td>
				378	<td>w1 w2 -- (w2<w1?w2:w1)</td>
				379	<td>Two values are popped off the stack. The larger one is pushed back
				380	onto the stack.</td>
				381	</tr>
				382	<tr><td>MAX</td>
				383	<td>MAX</td>
				384	<td>w1 w2 -- (w2>w1?w2:w1)</td>
				385	<td>Two values are popped off the stack. The larger value is pushed back
				386	onto the stack.</td>
				387	</tr>
				388	<tr><td colspan="4">STACK MANIPULATION OPERATIONS</td></tr>
				389	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				390	<tr><td>DROP</td>
				391	<td>DROP</td>
				392	<td>w -- </td>
				393	<td>One value is popped off the stack.</td>
				394	</tr>
				395	<tr><td>DROP2</td>
				396	<td>DROP2</td>
				397	<td>w1 w2 -- </td>
				398	<td>Two values are popped off the stack.</td>
				399	</tr>
				400	<tr><td>NIP</td>
				401	<td>NIP</td>
				402	<td>w1 w2 -- w2</td>
				403	<td>The second value on the stack is removed from the stack. That is,
				404	a value is popped off the stack and retained. Then a second value is
				405	popped and the retained value is pushed.</td>
				406	</tr>
				407	<tr><td>NIP2</td>
				408	<td>NIP2</td>
				409	<td>w1 w2 w3 w4 -- w3 w4</td>
				410	<td>The third and fourth values on the stack are removed from it. That is,
				411	two values are popped and retained. Then two more values are popped and
				412	the two retained values are pushed back on.</td>
				413	</tr>
				414	<tr><td>DUP</td>
				415	<td>DUP</td>
				416	<td>w1 -- w1 w1</td>
				417	<td>One value is popped off the stack. That value is then pushed onto
				418	the stack twice to duplicate the top stack vaue.</td>
				419	</tr>
				420	<tr><td>DUP2</td>
				421	<td>DUP2</td>
				422	<td>w1 w2 -- w1 w2 w1 w2</td>
				423	<td>The top two values on the stack are duplicated. That is, two vaues
				424	are popped off the stack. They are alternately pushed back on the
				425	stack twice each.</td>
				426	</tr>
				427	<tr><td>SWAP</td>
				428	<td>SWAP</td>
				429	<td>w1 w2 -- w2 w1</td>
				430	<td>The top two stack items are reversed in their order. That is, two
				431	values are popped off the stack and pushed back onto the stack in
				432	the opposite order they were popped.</td>
				433	</tr>
				434	<tr><td>SWAP2</td>
				435	<td>SWAP2</td>
				436	<td>w1 w2 w3 w4 -- w3 w4 w2 w1</td>
				437	<td>The top four stack items are swapped in pairs. That is, two values
				438	are popped and retained. Then, two more values are popped and retained.
				439	The values are pushed back onto the stack in the reverse order but
				440	in pairs.</p>
				441	</tr>
				442	<tr><td>OVER</td>
				443	<td>OVER</td>
				444	<td>w1 w2-- w1 w2 w1</td>
				445	<td>Two values are popped from the stack. They are pushed back
				446	onto the stack in the order w1 w2 w1. This seems to cause the
				447	top stack element to be duplicated "over" the next value.</td>
				448	</tr>
				449	<tr><td>OVER2</td>
				450	<td>OVER2</td>
				451	<td>w1 w2 w3 w4 -- w1 w2 w3 w4 w1 w2</td>
				452	<td>The third and fourth values on the stack are replicated onto the
				453	top of the stack</td>
				454	</tr>
				455	<tr><td>ROT</td>
				456	<td>ROT</td>
				457	<td>w1 w2 w3 -- w2 w3 w1</td>
				458	<td>The top three values are rotated. That is, three value are popped
				459	off the stack. They are pushed back onto the stack in the order
				460	w1 w3 w2.</td>
				461	</tr>
				462	<tr><td>ROT2</td>
				463	<td>ROT2</td>
				464	<td>w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2</td>
				465	<td>Like ROT but the rotation is done using three pairs instead of
				466	three singles.</td>
				467	</tr>
				468	<tr><td>RROT</td>
				469	<td>RROT</td>
				470	<td>w1 w2 w3 -- w2 w3 w1</td>
				471	<td>Reverse rotation. Like ROT, but it rotates the other way around.
				472	Essentially, the third element on the stack is moved to the top
				473	of the stack.</td>
				474	</tr>
				475	<tr><td>RROT2</td>
				476	<td>RROT2</td>
				477	<td>w1 w2 w3 w4 w5 w6 -- w3 w4 w5 w6 w1 w2</td>
				478	<td>Double reverse rotation. Like RROT but the rotation is done using
				479	three pairs instead of three singles. The fifth and sixth stack
				480	elements are moved to the first and second positions</td>
				481	</tr>
				482	<tr><td>TUCK</td>
				483	<td>TUCK</td>
				484	<td>w1 w2 -- w2 w1 w2</td>
				485	<td>Similar to OVER except that the second operand is being
				486	replicated. Essentially, the first operand is being "tucked"
				487	in between two instances of the second operand. Logically, two
				488	values are popped off the stack. They are placed back on the
				489	stack in the order w2 w1 w2.</td>
				490	</tr>
				491	<tr><td>TUCK2</td>
				492	<td>TUCK2</td>
				493	<td>w1 w2 w3 w4 -- w3 w4 w1 w2 w3 w4</td>
				494	<td>Like TUCK but a pair of elements is tucked over two pairs.
				495	That is, the top two elements of the stack are duplicated and
				496	inserted into the stack at the fifth and positions.</td>
				497	</tr>
				498	<tr><td>PICK</td>
				499	<td>PICK</td>
				500	<td>x0 ... Xn n -- x0 ... Xn x0</td>
				501	<td>The top of the stack is used as an index into the remainder of
				502	the stack. The element at the nth position replaces the index
				503	(top of stack). This is useful for cycling through a set of
				504	values. Note that indexing is zero based. So, if n=0 then you
				505	get the second item on the stack. If n=1 you get the third, etc.
				506	Note also that the index is replaced by the n'th value. </td>
				507	</tr>
				508	<tr><td>SELECT</td>
				509	<td>SELECT</td>
				510	<td>m n X0..Xm Xm+1 .. Xn -- Xm</td>
				511	<td>This is like PICK but the list is removed and you need to specify
				512	both the index and the size of the list. Careful with this one,
				513	the wrong value for n can blow away a huge amount of the stack.</td>
				514	</tr>
				515	<tr><td>ROLL</td>
				516	<td>ROLL</td>
				517	<td>x0 x1 .. xn n -- x1 .. xn x0</td>
				518	<td><b>Not Implemented</b>. This one has been left as an exercise to
				519	the student. If you can implement this one you understand Stacker
				520	and probably a fair amount about LLVM since this is one of the
				521	more complicated Stacker operations. See the StackerCompiler.cpp
				522	file in the projects/Stacker/lib/compiler directory. The operation
				523	of ROLL is like a generalized ROT. That is ROLL with n=1 is the
				524	same as ROT. The n value (top of stack) is used as an index to
				525	select a value up the stack that is <em>moved</em> to the top of
				526	the stack. See the implementations of PICk and SELECT to get
				527	some hints.<p>
				528	</tr>
				529	<tr><td colspan="4">MEMORY OPERATIONS</td></tr>
				530	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				531	<tr><td>MALLOC</td>
				532	<td>MALLOC</td>
				533	<td>w1 -- p</td>
				534	<td>One value is popped off the stack. The value is used as the size
				535	of a memory block to allocate. The size is in bytes, not words.
				536	The memory allocation is completed and the address of the memory
				537	block is pushed onto the stack.</td>
				538	</tr>
				539	<tr><td>FREE</td>
				540	<td>FREE</td>
				541	<td>p -- </td>
				542	<td>One pointer value is popped off the stack. The value should be
				543	the address of a memory block created by the MALLOC operation. The
				544	associated memory block is freed. Nothing is pushed back on the
				545	stack. Many bugs can be created by attempting to FREE something
				546	that isn't a pointer to a MALLOC allocated memory block. Make
				547	sure you know what's on the stack. One way to do this is with
				548	the following idiom:<br/>
				549	<code>64 MALLOC DUP DUP (use ptr) DUP (use ptr) ... FREE</code>
				550	<br/>This ensures that an extra copy of the pointer is placed on
				551	the stack (for the FREE at the end) and that every use of the
				552	pointer is preceded by a DUP to retain the copy for FREE.</td>
				553	</tr>
				554	<tr><td>GET</td>
				555	<td>GET</td>
				556	<td>w1 p -- w2 p</td>
				557	<td>An integer index and a pointer to a memory block are popped of
				558	the block. The index is used to index one byte from the memory
				559	block. That byte value is retained, the pointer is pushed again
				560	and the retained value is pushed. Note that the pointer value
				561	s essentially retained in its position so this doesn't count
				562	as a "use ptr" in the FREE idiom.</td>
				563	</tr>
				564	<tr><td>PUT</td>
				565	<td>PUT</td>
				566	<td>w1 w2 p -- p </td>
				567	<td>An integer value is popped of the stack. This is the value to
				568	be put into a memory block. Another integer value is popped of
				569	the stack. This is the indexed byte in the memory block. A
				570	pointer to the memory block is popped off the stack. The
				571	first value (w1) is then converted to a byte and written
				572	to the element of the memory block(p) at the index given
				573	by the second value (w2). The pointer to the memory block is
				574	pushed back on the stack so this doesn't count as a "use ptr"
				575	in the FREE idiom.</td>
				576	</tr>
				577	<tr><td colspan="4">CONTROL FLOW OPERATIONS</td></tr>
				578	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				579	<tr><td>RETURN</td>
				580	<td>RETURN</td>
				581	<td> -- </td>
				582	<td>The currently executing definition returns immediately to its caller.
				583	Note that there is an implicit <code>RETURN</code> at the end of each
				584	definition, logically located at the semi-colon. The sequence
				585	<code>RETURN ;</code> is valid but redundant.</td>
				586	</tr>
				587	<tr><td>EXIT</td>
				588	<td>EXIT</td>
				589	<td>w1 -- </td>
				590	<td>A return value for the program is popped off the stack. The program is
				591	then immediately terminated. This is normally an abnormal exit from the
				592	program. For a normal exit (when <code>MAIN</code> finishes), the exit
				593	code will always be zero in accordance with UNIX conventions.</td>
				594	</tr>
				595	<tr><td>RECURSE</td>
				596	<td>RECURSE</td>
				597	<td> -- </td>
				598	<td>The currently executed definition is called again. This operation is
				599	needed since the definition of a word doesn't exist until the semi colon
				600	is reacher. Attempting something like:<br/>
				601	<code> : recurser recurser ; </code><br/> will yield and error saying that
				602	"recurser" is not defined yet. To accomplish the same thing, change this
				603	to:<br/>
				604	<code> : recurser RECURSE ; </code></td>
				605	</tr>
				606	<tr><td>IF (words...) ENDIF</td>
				607	<td>IF (words...) ENDIF</td>
				608	<td>b -- </td>
				609	<td>A boolean value is popped of the stack. If it is non-zero then the "words..."
				610	are executed. Otherwise, execution continues immediately following the ENDIF.</td>
				611	</tr>
				612	<tr><td>IF (words...) ELSE (words...) ENDIF</td>
				613	<td>IF (words...) ELSE (words...) ENDIF</td>
				614	<td>b -- </td>
				615	<td>A boolean value is popped of the stack. If it is non-zero then the "words..."
				616	between IF and ELSE are executed. Otherwise the words between ELSE and ENDIF are
				617	executed. In either case, after the (words....) have executed, execution continues
				618	immediately following the ENDIF. </td>
				619	</tr>
				620	<tr><td>WHILE (words...) END</td>
				621	<td>WHILE (words...) END</td>
				622	<td>b -- b </td>
				623	<td>The boolean value on the top of the stack is examined. If it is non-zero then the
				624	"words..." between WHILE and END are executed. Execution then begins again at the WHILE where another
				625	boolean is popped off the stack. To prevent this operation from eating up the entire
				626	stack, you should push onto the stack (just before the END) a boolean value that indicates
				627	whether to terminate. Note that since booleans and integers can be coerced you can
				628	use the following "for loop" idiom:<br/>
				629	<code>(push count) WHILE (words...) -- END</code><br/>
				630	For example:<br/>
				631	<code>10 WHILE DUP >d -- END</code><br/>
				632	This will print the numbers from 10 down to 1. 10 is pushed on the stack. Since that is
				633	non-zero, the while loop is entered. The top of the stack (10) is duplicated and then
				634	printed out with >d. The top of the stack is decremented, yielding 9 and control is
				635	transfered back to the WHILE keyword. The process starts all over again and repeats until
				636	the top of stack is decremented to 0 at which the WHILE test fails and control is
				637	transfered to the word after the END.</td>
				638	</tr>
				639	<tr><td colspan="4">INPUT & OUTPUT OPERATIONS</td></tr>
				640	<tr><td>Word</td><td>Name</td><td>Operation</td><td>Description</td></tr>
				641	<tr><td>SPACE</td>
				642	<td>SPACE</td>
				643	<td> -- </td>
				644	<td>A space character is put out. There is no stack effect.</td>
				645	</tr>
				646	<tr><td>TAB</td>
				647	<td>TAB</td>
				648	<td> -- </td>
				649	<td>A tab character is put out. There is no stack effect.</td>
				650	</tr>
				651	<tr><td>CR</td>
				652	<td>CR</td>
				653	<td> -- </td>
				654	<td>A carriage return character is put out. There is no stack effect.</td>
				655	</tr>
				656	<tr><td>>s</td>
				657	<td>OUT_STR</td>
				658	<td> -- </td>
				659	<td>A string pointer is popped from the stack. It is put out.</td>
				660	</tr>
				661	<tr><td>>d</td>
				662	<td>OUT_STR</td>
				663	<td> -- </td>
				664	<td>A value is popped from the stack. It is put out as a decimal integer.</td>
				665	</tr>
				666	<tr><td>>c</td>
				667	<td>OUT_CHR</td>
				668	<td> -- </td>
				669	<td>A value is popped from the stack. It is put out as an ASCII character.</td>
				670	</tr>
				671	<tr><td><s</td>
				672	<td>IN_STR</td>
				673	<td> -- s </td>
				674	<td>A string is read from the input via the scanf(3) format string " %as". The
				675	resulting string is pushed onto the stack.</td>
				676	</tr>
				677	<tr><td><d</td>
				678	<td>IN_STR</td>
				679	<td> -- w </td>
				680	<td>An integer is read from the input via the scanf(3) format string " %d". The
				681	resulting value is pushed onto the stack</td>
				682	</tr>
				683	<tr><td><c</td>
				684	<td>IN_CHR</td>
				685	<td> -- w </td>
				686	<td>A single character is read from the input via the scanf(3) format string
				687	" %c". The value is converted to an integer and pushed onto the stack.</td>
				688	</tr>
				689	<tr><td>DUMP</td>
				690	<td>DUMP</td>
				691	<td> -- </td>
				692	<td>The stack contents are dumped to standard output. This is useful for
				693	debugging your definitions. Put DUMP at the beginning and end of a definition
				694	to see instantly the net effect of the definition.</td>
				695	</tr>
				696	</table>
				697	</div>
				698	<!-- ======================================================================= -->
				699	<div class="doc_section"> <a name="directory">Directory Structure</a></div>
				700	<div class="doc_text">
				701	<p>The source code, test programs, and sample programs can all be found
				702	under the LLVM "projects" directory. You will need to obtain the LLVM sources
				703	to find it (either via anonymous CVS or a tarball. See the
				704	<a href="GettingStarted.html">Getting Started</a> document).</p>
				705	<p>Under the "projects" directory there is a directory named "stacker". That
				706	directory contains everything, as follows:</p>
				707	<ul>
				708	<li><em>lib</em> - contains most of the source code
				709	<ul>
				710	<li><em>lib/compiler</em> - contains the compiler library
				711	<li><em>lib/runtime</em> - contains the runtime library
				712	</ul></li>
				713	<li><em>test</em> - contains the test programs</li>
				714	<li><em>tools</em> - contains the Stacker compiler main program, stkrc
				715	<ul>
				716	<li><em>lib/stkrc</em> - contains the Stacker compiler main program
				717	</ul</li>
				718	<li><em>sample</em> - contains the sample programs</li>
				719	</ul>
				720	</div>
				721	<!-- ======================================================================= -->
				722	<div class="doc_section"> <a name="directory">Prime: A Complete Example</a></div>
				723	<div class="doc_text">
				724	<p>The following fully documented program highlights many of features of both
				725	the Stacker language and what is possible with LLVM. The program simply
				726	prints out the prime numbers until it reaches
				727	</p>
				728	</div>
				729	<div class="doc_text">
				730	<p><code>
				731	<![CDATA[
				732	################################################################################
				733	#
				734	# Brute force prime number generator
				735	#
				736	# This program is written in classic Stacker style, that being the style of a
				737	# stack. Start at the bottom and read your way up !
				738	#
				739	# Reid Spencer - Nov 2003
				740	################################################################################
				741	# Utility definitions
				742	################################################################################
				743	: print >d CR ;
				744	: it_is_a_prime TRUE ;
				745	: it_is_not_a_prime FALSE ;
				746	: continue_loop TRUE ;
				747	: exit_loop FALSE;
				748
				749	################################################################################
				750	# This definition tryies an actual division of a candidate prime number. It
				751	# determines whether the division loop on this candidate should continue or
				752	# not.
				753	# STACK<:
				754	# div - the divisor to try
				755	# p - the prime number we are working on
				756	# STACK>:
				757	# cont - should we continue the loop ?
				758	# div - the next divisor to try
				759	# p - the prime number we are working on
				760	################################################################################
				761	: try_dividing
				762	DUP2 ( save div and p )
				763	SWAP ( swap to put divisor second on stack)
				764	MOD 0 = ( get remainder after division and test for 0 )
				765	IF
				766	exit_loop ( remainder = 0, time to exit )
				767	ELSE
				768	continue_loop ( remainder != 0, keep going )
				769	ENDIF
				770	;
				771
				772	################################################################################
				773	# This function tries one divisor by calling try_dividing. But, before doing
				774	# that it checks to see if the value is 1. If it is, it does not bother with
				775	# the division because prime numbers are allowed to be divided by one. The
				776	# top stack value (cont) is set to determine if the loop should continue on
				777	# this prime number or not.
				778	# STACK<:
				779	# cont - should we continue the loop (ignored)?
				780	# div - the divisor to try
				781	# p - the prime number we are working on
				782	# STACK>:
				783	# cont - should we continue the loop ?
				784	# div - the next divisor to try
				785	# p - the prime number we are working on
				786	################################################################################
				787	: try_one_divisor
				788	DROP ( drop the loop continuation )
				789	DUP ( save the divisor )
				790	1 = IF ( see if divisor is == 1 )
				791	exit_loop ( no point dividing by 1 )
				792	ELSE
				793	try_dividing ( have to keep going )
				794	ENDIF
				795	SWAP ( get divisor on top )
				796	-- ( decrement it )
				797	SWAP ( put loop continuation back on top )
				798	;
				799
				800	################################################################################
				801	# The number on the stack (p) is a candidate prime number that we must test to
				802	# determine if it really is a prime number. To do this, we divide it by every
				803	# number from one p-1 to 1. The division is handled in the try_one_divisor
				804	# definition which returns a loop continuation value (which we also seed with
				805	# the value 1). After the loop, we check the divisor. If it decremented all
				806	# the way to zero then we found a prime, otherwise we did not find one.
				807	# STACK<:
				808	# p - the prime number to check
				809	# STACK>:
				810	# yn - boolean indiating if its a prime or not
				811	# p - the prime number checked
				812	################################################################################
				813	: try_harder
				814	DUP ( duplicate to get divisor value ) )
				815	-- ( first divisor is one less than p )
				816	1 ( continue the loop )
				817	WHILE
				818	try_one_divisor ( see if its prime )
				819	END
				820	DROP ( drop the continuation value )
				821	0 = IF ( test for divisor == 1 )
				822	it_is_a_prime ( we found one )
				823	ELSE
				824	it_is_not_a_prime ( nope, this one is not a prime )
				825	ENDIF
				826	;
				827
				828	################################################################################
				829	# This definition determines if the number on the top of the stack is a prime
				830	# or not. It does this by testing if the value is degenerate (<= 3) and
				831	# responding with yes, its a prime. Otherwise, it calls try_harder to actually
				832	# make some calculations to determine its primeness.
				833	# STACK<:
				834	# p - the prime number to check
				835	# STACK>:
				836	# yn - boolean indicating if its a prime or not
				837	# p - the prime number checked
				838	################################################################################
				839	: is_prime
				840	DUP ( save the prime number )
				841	3 >= IF ( see if its <= 3 )
				842	it_is_a_prime ( its <= 3 just indicate its prime )
				843	ELSE
				844	try_harder ( have to do a little more work )
				845	ENDIF
				846	;
				847
				848	################################################################################
				849	# This definition is called when it is time to exit the program, after we have
				850	# found a sufficiently large number of primes.
				851	# STACK<: ignored
				852	# STACK>: exits
				853	################################################################################
				854	: done
				855	"Finished" >s CR ( say we are finished )
				856	0 EXIT ( exit nicely )
				857	;
				858
				859	################################################################################
				860	# This definition checks to see if the candidate is greater than the limit. If
				861	# it is, it terminates the program by calling done. Otherwise, it increments
				862	# the value and calls is_prime to determine if the candidate is a prime or not.
				863	# If it is a prime, it prints it. Note that the boolean result from is_prime is
				864	# gobbled by the following IF which returns the stack to just contining the
				865	# prime number just considered.
				866	# STACK<:
				867	# p - one less than the prime number to consider
				868	# STACK>
				869	# p+1 - the prime number considered
				870	################################################################################
				871	: consider_prime
				872	DUP ( save the prime number to consider )
				873	1000000 < IF ( check to see if we are done yet )
				874	done ( we are done, call "done" )
				875	ENDIF
				876	++ ( increment to next prime number )
				877	is_prime ( see if it is a prime )
				878	IF
				879	print ( it is, print it )
				880	ENDIF
				881	;
				882
				883	################################################################################
				884	# This definition starts at one, prints it out and continues into a loop calling
				885	# consider_prime on each iteration. The prime number candidate we are looking at
				886	# is incremented by consider_prime.
				887	# STACK<: empty
				888	# STACK>: empty
				889	################################################################################
				890	: find_primes
				891	"Prime Numbers: " >s CR ( say hello )
				892	DROP ( get rid of that pesky string )
				893	1 ( stoke the fires )
				894	print ( print the first one, we know its prime )
				895	WHILE ( loop while the prime to consider is non zero )
				896	consider_prime ( consider one prime number )
				897	END
				898	;
				899
				900	################################################################################
				901	#
				902	################################################################################
				903	: say_yes
				904	>d ( Print the prime number )
				905	" is prime." ( push string to output )
				906	>s ( output it )
				907	CR ( print carriage return )
				908	DROP ( pop string )
				909	;
				910
				911	: say_no
				912	>d ( Print the prime number )
				913	" is NOT prime." ( push string to put out )
				914	>s ( put out the string )
				915	CR ( print carriage return )
				916	DROP ( pop string )
				917	;
				918
				919	################################################################################
				920	# This definition processes a single command line argument and determines if it
				921	# is a prime number or not.
				922	# STACK<:
				923	# n - number of arguments
				924	# arg1 - the prime numbers to examine
				925	# STACK>:
				926	# n-1 - one less than number of arguments
				927	# arg2 - we processed one argument
				928	################################################################################
				929	: do_one_argument
				930	-- ( decrement loop counter )
				931	SWAP ( get the argument value )
				932	is_prime IF ( determine if its prime )
				933	say_yes ( uhuh )
				934	ELSE
				935	say_no ( nope )
				936	ENDIF
				937	DROP ( done with that argument )
				938	;
				939
				940	################################################################################
				941	# The MAIN program just prints a banner and processes its arguments.
				942	# STACK<:
				943	# n - number of arguments
				944	# ... - the arguments
				945	################################################################################
				946	: process_arguments
				947	WHILE ( while there are more arguments )
				948	do_one_argument ( process one argument )
				949	END
				950	;
				951
				952	################################################################################
				953	# The MAIN program just prints a banner and processes its arguments.
				954	# STACK<: arguments
				955	################################################################################
				956	: MAIN
				957	NIP ( get rid of the program name )
				958	-- ( reduce number of arguments )
				959	DUP ( save the arg counter )
				960	1 <= IF ( See if we got an argument )
				961	process_arguments ( tell user if they are prime )
				962	ELSE
				963	find_primes ( see how many we can find )
				964	ENDIF
				965	0 ( push return code )
				966	;
				967	]]>
				968	</code>
				969	</p>
				970	</div>
				971	<!-- ======================================================================= -->
				972	<div class="doc_section"> <a name="lexicon">Internals</a></div>
				973	<div class="doc_text"><p>To be completed.</p></div>
				974	<div class="doc_subsection"><a name="stack"></a>The Lexer</div>
				975	<div class="doc_subsection"><a name="stack"></a>The Parser</div>
				976	<div class="doc_subsection"><a name="stack"></a>The Compiler</div>
				977	<div class="doc_subsection"><a name="stack"></a>The Stack</div>
				978	<div class="doc_subsection"><a name="stack"></a>Definitions Are Functions</div>
				979	<div class="doc_subsection"><a name="stack"></a>Words Are BasicBlocks</div>
				980	<!-- ======================================================================= -->
				981	<hr>
				982	<div class="doc_footer">
				983	<address><a href="mailto:rspencer@x10sys.com">Reid Spencer</a></address>
				984	<a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a>
				985	<br>Last modified: $Date$ </div>
				986	</body>
				987	</html>