001 /* 002 * Licensed to the Apache Software Foundation (ASF) under one 003 * or more contributor license agreements. See the NOTICE file 004 * distributed with this work for additional information 005 * regarding copyright ownership. The ASF licenses this file 006 * to you under the Apache License, Version 2.0 (the "License"); 007 * you may not use this file except in compliance with the License. 008 * You may obtain a copy of the License at 009 * 010 * http://www.apache.org/licenses/LICENSE-2.0 011 * 012 * Unless required by applicable law or agreed to in writing, software 013 * distributed under the License is distributed on an "AS IS" BASIS, 014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 015 * See the License for the specific language governing permissions and 016 * limitations under the License. 017 */ 018 /* 019 * $Id: CoroutineManager.java 468653 2006-10-28 07:07:05Z minchau $ 020 */ 021 package org.apache.xml.dtm.ref; 022 023 import java.util.BitSet; 024 025 import org.apache.xml.res.XMLErrorResources; 026 import org.apache.xml.res.XMLMessages; 027 028 029 /** 030 * <p>Support the coroutine design pattern.</p> 031 * 032 * <p>A coroutine set is a very simple cooperative non-preemptive 033 * multitasking model, where the switch from one task to another is 034 * performed via an explicit request. Coroutines interact according to 035 * the following rules:</p> 036 * 037 * <ul> 038 * <li>One coroutine in the set has control, which it retains until it 039 * either exits or resumes another coroutine.</li> 040 * <li>A coroutine is activated when it is resumed by some other coroutine 041 * for the first time.</li> 042 * <li>An active coroutine that gives up control by resuming another in 043 * the set retains its context -- including call stack and local variables 044 * -- so that if/when it is resumed, it will proceed from the point at which 045 * it last gave up control.</li> 046 * </ul> 047 * 048 * <p>Coroutines can be thought of as falling somewhere between pipes and 049 * subroutines. Like call/return, there is an explicit flow of control 050 * from one coroutine to another. Like pipes, neither coroutine is 051 * actually "in charge", and neither must exit in order to transfer 052 * control to the other. </p> 053 * 054 * <p>One classic application of coroutines is in compilers, where both 055 * the parser and the lexer are maintaining complex state 056 * information. The parser resumes the lexer to process incoming 057 * characters into lexical tokens, and the lexer resumes the parser 058 * when it has reached a point at which it has a reliably interpreted 059 * set of tokens available for semantic processing. Structuring this 060 * as call-and-return would require saving and restoring a 061 * considerable amount of state each time. Structuring it as two tasks 062 * connected by a queue may involve higher overhead (in systems which 063 * can optimize the coroutine metaphor), isn't necessarily as clear in 064 * intent, may have trouble handling cases where data flows in both 065 * directions, and may not handle some of the more complex cases where 066 * more than two coroutines are involved.</p> 067 * 068 * <p>Most coroutine systems also provide a way to pass data between the 069 * source and target of a resume operation; this is sometimes referred 070 * to as "yielding" a value. Others rely on the fact that, since only 071 * one member of a coroutine set is running at a time and does not 072 * lose control until it chooses to do so, data structures may be 073 * directly shared between them with only minimal precautions.</p> 074 * 075 * <p>"Note: This should not be taken to mean that producer/consumer 076 * problems should be always be done with coroutines." Queueing is 077 * often a better solution when only two threads of execution are 078 * involved and full two-way handshaking is not required. It's a bit 079 * difficult to find short pedagogical examples that require 080 * coroutines for a clear solution.</p> 081 * 082 * <p>The fact that only one of a group of coroutines is running at a 083 * time, and the control transfer between them is explicit, simplifies 084 * their possible interactions, and in some implementations permits 085 * them to be implemented more efficiently than general multitasking. 086 * In some situations, coroutines can be compiled out entirely; 087 * in others, they may only require a few instructions more than a 088 * simple function call.</p> 089 * 090 * <p>This version is built on top of standard Java threading, since 091 * that's all we have available right now. It's been encapsulated for 092 * code clarity and possible future optimization.</p> 093 * 094 * <p>(Two possible approaches: wait-notify based and queue-based. Some 095 * folks think that a one-item queue is a cleaner solution because it's 096 * more abstract -- but since coroutine _is_ an abstraction I'm not really 097 * worried about that; folks should be able to switch this code without 098 * concern.)</p> 099 * 100 * <p>%TBD% THIS SHOULD BE AN INTERFACE, to facilitate building other 101 * implementations... perhaps including a true coroutine system 102 * someday, rather than controlled threading. Arguably Coroutine 103 * itself should be an interface much like Runnable, but I think that 104 * can be built on top of this.</p> 105 * */ 106 public class CoroutineManager 107 { 108 /** "Is this coroutine ID number already in use" lookup table. 109 * Currently implemented as a bitset as a compromise between 110 * compactness and speed of access, but obviously other solutions 111 * could be applied. 112 * */ 113 BitSet m_activeIDs=new BitSet(); 114 115 /** Limit on the coroutine ID numbers accepted. I didn't want the 116 * in-use table to grow without bound. If we switch to a more efficient 117 * sparse-array mechanism, it may be possible to raise or eliminate 118 * this boundary. 119 * @xsl.usage internal 120 */ 121 static final int m_unreasonableId=1024; 122 123 /** Internal field used to hold the data being explicitly passed 124 * from one coroutine to another during a co_resume() operation. 125 * (Of course implicit data sharing may also occur; one of the reasons 126 * for using coroutines is that you're guaranteed that none of the 127 * other coroutines in your set are using shared structures at the time 128 * you access them.) 129 * 130 * %REVIEW% It's been proposed that we be able to pass types of data 131 * other than Object -- more specific object types, or 132 * lighter-weight primitives. This would seem to create a potential 133 * explosion of "pass x recieve y back" methods (or require 134 * fracturing resume into two calls, resume-other and 135 * wait-to-be-resumed), and the weight issue could be managed by 136 * reusing a mutable buffer object to contain the primitive 137 * (remember that only one coroutine runs at a time, so once the 138 * buffer's set it won't be walked on). Typechecking objects is 139 * interesting from a code-robustness point of view, but it's 140 * unclear whether it makes sense to encapsulate that in the 141 * coroutine code or let the callers do it, since it depends on RTTI 142 * either way. Restricting the parameters to objects implementing a 143 * specific CoroutineParameter interface does _not_ seem to be a net 144 * win; applications can do so if they want via front-end code, but 145 * there seem to be too many use cases involving passing an existing 146 * object type that you may not have the freedom to alter and may 147 * not want to spend time wrapping another object around. 148 * */ 149 Object m_yield=null; 150 151 // Expose??? 152 final static int NOBODY=-1; 153 final static int ANYBODY=-1; 154 155 /** Internal field used to confirm that the coroutine now waking up is 156 * in fact the one we intended to resume. Some such selection mechanism 157 * is needed when more that two coroutines are operating within the same 158 * group. 159 */ 160 int m_nextCoroutine=NOBODY; 161 162 /** <p>Each coroutine in the set managed by a single 163 * CoroutineManager is identified by a small positive integer. This 164 * brings up the question of how to manage those integers to avoid 165 * reuse... since if two coroutines use the same ID number, resuming 166 * that ID could resume either. I can see arguments for either 167 * allowing applications to select their own numbers (they may want 168 * to declare mnemonics via manefest constants) or generating 169 * numbers on demand. This routine's intended to support both 170 * approaches.</p> 171 * 172 * <p>%REVIEW% We could use an object as the identifier. Not sure 173 * it's a net gain, though it would allow the thread to be its own 174 * ID. Ponder.</p> 175 * 176 * @param coroutineID If >=0, requests that we reserve this number. 177 * If <0, requests that we find, reserve, and return an available ID 178 * number. 179 * 180 * @return If >=0, the ID number to be used by this coroutine. If <0, 181 * an error occurred -- the ID requested was already in use, or we 182 * couldn't assign one without going over the "unreasonable value" mark 183 * */ 184 public synchronized int co_joinCoroutineSet(int coroutineID) 185 { 186 if(coroutineID>=0) 187 { 188 if(coroutineID>=m_unreasonableId || m_activeIDs.get(coroutineID)) 189 return -1; 190 } 191 else 192 { 193 // What I want is "Find first clear bit". That doesn't exist. 194 // JDK1.2 added "find last set bit", but that doesn't help now. 195 coroutineID=0; 196 while(coroutineID<m_unreasonableId) 197 { 198 if(m_activeIDs.get(coroutineID)) 199 ++coroutineID; 200 else 201 break; 202 } 203 if(coroutineID>=m_unreasonableId) 204 return -1; 205 } 206 207 m_activeIDs.set(coroutineID); 208 return coroutineID; 209 } 210 211 /** In the standard coroutine architecture, coroutines are 212 * identified by their method names and are launched and run up to 213 * their first yield by simply resuming them; its's presumed that 214 * this recognizes the not-already-running case and does the right 215 * thing. We seem to need a way to achieve that same threadsafe 216 * run-up... eg, start the coroutine with a wait. 217 * 218 * %TBD% whether this makes any sense... 219 * 220 * @param thisCoroutine the identifier of this coroutine, so we can 221 * recognize when we are being resumed. 222 * @exception java.lang.NoSuchMethodException if thisCoroutine isn't 223 * a registered member of this group. %REVIEW% whether this is the 224 * best choice. 225 * */ 226 public synchronized Object co_entry_pause(int thisCoroutine) throws java.lang.NoSuchMethodException 227 { 228 if(!m_activeIDs.get(thisCoroutine)) 229 throw new java.lang.NoSuchMethodException(); 230 231 while(m_nextCoroutine != thisCoroutine) 232 { 233 try 234 { 235 wait(); 236 } 237 catch(java.lang.InterruptedException e) 238 { 239 // %TBD% -- Declare? Encapsulate? Ignore? Or 240 // dance widdershins about the instruction cache? 241 } 242 } 243 244 return m_yield; 245 } 246 247 /** Transfer control to another coroutine which has already been started and 248 * is waiting on this CoroutineManager. We won't return from this call 249 * until that routine has relinquished control. 250 * 251 * %TBD% What should we do if toCoroutine isn't registered? Exception? 252 * 253 * @param arg_object A value to be passed to the other coroutine. 254 * @param thisCoroutine Integer identifier for this coroutine. This is the 255 * ID we watch for to see if we're the ones being resumed. 256 * @param toCoroutine Integer identifier for the coroutine we wish to 257 * invoke. 258 * @exception java.lang.NoSuchMethodException if toCoroutine isn't a 259 * registered member of this group. %REVIEW% whether this is the best choice. 260 * */ 261 public synchronized Object co_resume(Object arg_object,int thisCoroutine,int toCoroutine) throws java.lang.NoSuchMethodException 262 { 263 if(!m_activeIDs.get(toCoroutine)) 264 throw new java.lang.NoSuchMethodException(XMLMessages.createXMLMessage(XMLErrorResources.ER_COROUTINE_NOT_AVAIL, new Object[]{Integer.toString(toCoroutine)})); //"Coroutine not available, id="+toCoroutine); 265 266 // We expect these values to be overwritten during the notify()/wait() 267 // periods, as other coroutines in this set get their opportunity to run. 268 m_yield=arg_object; 269 m_nextCoroutine=toCoroutine; 270 271 notify(); 272 while(m_nextCoroutine != thisCoroutine || m_nextCoroutine==ANYBODY || m_nextCoroutine==NOBODY) 273 { 274 try 275 { 276 // System.out.println("waiting..."); 277 wait(); 278 } 279 catch(java.lang.InterruptedException e) 280 { 281 // %TBD% -- Declare? Encapsulate? Ignore? Or 282 // dance deasil about the program counter? 283 } 284 } 285 286 if(m_nextCoroutine==NOBODY) 287 { 288 // Pass it along 289 co_exit(thisCoroutine); 290 // And inform this coroutine that its partners are Going Away 291 // %REVIEW% Should this throw/return something more useful? 292 throw new java.lang.NoSuchMethodException(XMLMessages.createXMLMessage(XMLErrorResources.ER_COROUTINE_CO_EXIT, null)); //"CoroutineManager recieved co_exit() request"); 293 } 294 295 return m_yield; 296 } 297 298 /** Terminate this entire set of coroutines. The others will be 299 * deregistered and have exceptions thrown at them. Note that this 300 * is intended as a panic-shutdown operation; under normal 301 * circumstances a coroutine should always end with co_exit_to() in 302 * order to politely inform at least one of its partners that it is 303 * going away. 304 * 305 * %TBD% This may need significantly more work. 306 * 307 * %TBD% Should this just be co_exit_to(,,CoroutineManager.PANIC)? 308 * 309 * @param thisCoroutine Integer identifier for the coroutine requesting exit. 310 * */ 311 public synchronized void co_exit(int thisCoroutine) 312 { 313 m_activeIDs.clear(thisCoroutine); 314 m_nextCoroutine=NOBODY; // %REVIEW% 315 notify(); 316 } 317 318 /** Make the ID available for reuse and terminate this coroutine, 319 * transferring control to the specified coroutine. Note that this 320 * returns immediately rather than waiting for any further coroutine 321 * traffic, so the thread can proceed with other shutdown activities. 322 * 323 * @param arg_object A value to be passed to the other coroutine. 324 * @param thisCoroutine Integer identifier for the coroutine leaving the set. 325 * @param toCoroutine Integer identifier for the coroutine we wish to 326 * invoke. 327 * @exception java.lang.NoSuchMethodException if toCoroutine isn't a 328 * registered member of this group. %REVIEW% whether this is the best choice. 329 * */ 330 public synchronized void co_exit_to(Object arg_object,int thisCoroutine,int toCoroutine) throws java.lang.NoSuchMethodException 331 { 332 if(!m_activeIDs.get(toCoroutine)) 333 throw new java.lang.NoSuchMethodException(XMLMessages.createXMLMessage(XMLErrorResources.ER_COROUTINE_NOT_AVAIL, new Object[]{Integer.toString(toCoroutine)})); //"Coroutine not available, id="+toCoroutine); 334 335 // We expect these values to be overwritten during the notify()/wait() 336 // periods, as other coroutines in this set get their opportunity to run. 337 m_yield=arg_object; 338 m_nextCoroutine=toCoroutine; 339 340 m_activeIDs.clear(thisCoroutine); 341 342 notify(); 343 } 344 }