Tristan Matthews | 0a329cc | 2013-07-17 13:20:14 -0400 | [diff] [blame] | 1 | //------------------------------------------------------------------------------
|
| 2 | // File: WXList.cpp
|
| 3 | //
|
| 4 | // Desc: DirectShow base classes - implements a non-MFC based generic list
|
| 5 | // template class.
|
| 6 | // Copyright (c) 1992-2001 Microsoft Corporation. All rights reserved.
|
| 7 | //------------------------------------------------------------------------------
|
| 8 |
|
| 9 | #include <pjmedia-videodev/config.h>
|
| 10 |
|
| 11 | #if defined(PJMEDIA_VIDEO_DEV_HAS_DSHOW) && PJMEDIA_VIDEO_DEV_HAS_DSHOW != 0
|
| 12 |
|
| 13 | /* A generic list of pointers to objects.
|
| 14 | Objectives: avoid using MFC libraries in ndm kernel mode and
|
| 15 | provide a really useful list type.
|
| 16 |
|
| 17 | The class is thread safe in that separate threads may add and
|
| 18 | delete items in the list concurrently although the application
|
| 19 | must ensure that constructor and destructor access is suitably
|
| 20 | synchronised.
|
| 21 |
|
| 22 | The list name must not conflict with MFC classes as an
|
| 23 | application may use both
|
| 24 |
|
| 25 | The nodes form a doubly linked, NULL terminated chain with an anchor
|
| 26 | block (the list object per se) holding pointers to the first and last
|
| 27 | nodes and a count of the nodes.
|
| 28 | There is a node cache to reduce the allocation and freeing overhead.
|
| 29 | It optionally (determined at construction time) has an Event which is
|
| 30 | set whenever the list becomes non-empty and reset whenever it becomes
|
| 31 | empty.
|
| 32 | It optionally (determined at construction time) has a Critical Section
|
| 33 | which is entered during the important part of each operation. (About
|
| 34 | all you can do outside it is some parameter checking).
|
| 35 |
|
| 36 | The node cache is a repository of nodes that are NOT in the list to speed
|
| 37 | up storage allocation. Each list has its own cache to reduce locking and
|
| 38 | serialising. The list accesses are serialised anyway for a given list - a
|
| 39 | common cache would mean that we would have to separately serialise access
|
| 40 | of all lists within the cache. Because the cache only stores nodes that are
|
| 41 | not in the list, releasing the cache does not release any list nodes. This
|
| 42 | means that list nodes can be copied or rechained from one list to another
|
| 43 | without danger of creating a dangling reference if the original cache goes
|
| 44 | away.
|
| 45 |
|
| 46 | Questionable design decisions:
|
| 47 | 1. Retaining the warts for compatibility
|
| 48 | 2. Keeping an element count -i.e. counting whenever we do anything
|
| 49 | instead of only when we want the count.
|
| 50 | 3. Making the chain pointers NULL terminated. If the list object
|
| 51 | itself looks just like a node and the list is kept as a ring then
|
| 52 | it reduces the number of special cases. All inserts look the same.
|
| 53 | */
|
| 54 |
|
| 55 |
|
| 56 | #include <streams.h>
|
| 57 |
|
| 58 | /* set cursor to the position of each element of list in turn */
|
| 59 | #define INTERNALTRAVERSELIST(list, cursor) \
|
| 60 | for ( cursor = (list).GetHeadPositionI() \
|
| 61 | ; cursor!=NULL \
|
| 62 | ; cursor = (list).Next(cursor) \
|
| 63 | )
|
| 64 |
|
| 65 |
|
| 66 | /* set cursor to the position of each element of list in turn
|
| 67 | in reverse order
|
| 68 | */
|
| 69 | #define INTERNALREVERSETRAVERSELIST(list, cursor) \
|
| 70 | for ( cursor = (list).GetTailPositionI() \
|
| 71 | ; cursor!=NULL \
|
| 72 | ; cursor = (list).Prev(cursor) \
|
| 73 | )
|
| 74 |
|
| 75 | /* Constructor calls a separate initialisation function that
|
| 76 | creates a node cache, optionally creates a lock object
|
| 77 | and optionally creates a signaling object.
|
| 78 |
|
| 79 | By default we create a locking object, a DEFAULTCACHE sized
|
| 80 | cache but no event object so the list cannot be used in calls
|
| 81 | to WaitForSingleObject
|
| 82 | */
|
| 83 | CBaseList::CBaseList(__in_opt LPCTSTR pName, // Descriptive list name
|
| 84 | INT iItems) : // Node cache size
|
| 85 | #ifdef DEBUG
|
| 86 | CBaseObject(pName),
|
| 87 | #endif
|
| 88 | m_pFirst(NULL),
|
| 89 | m_pLast(NULL),
|
| 90 | m_Count(0),
|
| 91 | m_Cache(iItems)
|
| 92 | {
|
| 93 | } // constructor
|
| 94 |
|
| 95 | CBaseList::CBaseList(__in_opt LPCTSTR pName) : // Descriptive list name
|
| 96 | #ifdef DEBUG
|
| 97 | CBaseObject(pName),
|
| 98 | #endif
|
| 99 | m_pFirst(NULL),
|
| 100 | m_pLast(NULL),
|
| 101 | m_Count(0),
|
| 102 | m_Cache(DEFAULTCACHE)
|
| 103 | {
|
| 104 | } // constructor
|
| 105 |
|
| 106 | #ifdef UNICODE
|
| 107 | CBaseList::CBaseList(__in_opt LPCSTR pName, // Descriptive list name
|
| 108 | INT iItems) : // Node cache size
|
| 109 | #ifdef DEBUG
|
| 110 | CBaseObject(pName),
|
| 111 | #endif
|
| 112 | m_pFirst(NULL),
|
| 113 | m_pLast(NULL),
|
| 114 | m_Count(0),
|
| 115 | m_Cache(iItems)
|
| 116 | {
|
| 117 | } // constructor
|
| 118 |
|
| 119 | CBaseList::CBaseList(__in_opt LPCSTR pName) : // Descriptive list name
|
| 120 | #ifdef DEBUG
|
| 121 | CBaseObject(pName),
|
| 122 | #endif
|
| 123 | m_pFirst(NULL),
|
| 124 | m_pLast(NULL),
|
| 125 | m_Count(0),
|
| 126 | m_Cache(DEFAULTCACHE)
|
| 127 | {
|
| 128 | } // constructor
|
| 129 |
|
| 130 | #endif
|
| 131 |
|
| 132 | /* The destructor enumerates all the node objects in the list and
|
| 133 | in the cache deleting each in turn. We do not do any processing
|
| 134 | on the objects that the list holds (i.e. points to) so if they
|
| 135 | represent interfaces for example the creator of the list should
|
| 136 | ensure that each of them is released before deleting us
|
| 137 | */
|
| 138 | CBaseList::~CBaseList()
|
| 139 | {
|
| 140 | /* Delete all our list nodes */
|
| 141 |
|
| 142 | RemoveAll();
|
| 143 |
|
| 144 | } // destructor
|
| 145 |
|
| 146 | /* Remove all the nodes from the list but don't do anything
|
| 147 | with the objects that each node looks after (this is the
|
| 148 | responsibility of the creator).
|
| 149 | Aa a last act we reset the signalling event
|
| 150 | (if available) to indicate to clients that the list
|
| 151 | does not have any entries in it.
|
| 152 | */
|
| 153 | void CBaseList::RemoveAll()
|
| 154 | {
|
| 155 | /* Free up all the CNode objects NOTE we don't bother putting the
|
| 156 | deleted nodes into the cache as this method is only really called
|
| 157 | in serious times of change such as when we are being deleted at
|
| 158 | which point the cache will be deleted anway */
|
| 159 |
|
| 160 | CNode *pn = m_pFirst;
|
| 161 | while (pn) {
|
| 162 | CNode *op = pn;
|
| 163 | pn = pn->Next();
|
| 164 | delete op;
|
| 165 | }
|
| 166 |
|
| 167 | /* Reset the object count and the list pointers */
|
| 168 |
|
| 169 | m_Count = 0;
|
| 170 | m_pFirst = m_pLast = NULL;
|
| 171 |
|
| 172 | } // RemoveAll
|
| 173 |
|
| 174 |
|
| 175 |
|
| 176 | /* Return a position enumerator for the entire list.
|
| 177 | A position enumerator is a pointer to a node object cast to a
|
| 178 | transparent type so all we do is return the head/tail node
|
| 179 | pointer in the list.
|
| 180 | WARNING because the position is a pointer to a node there is
|
| 181 | an implicit assumption for users a the list class that after
|
| 182 | deleting an object from the list that any other position
|
| 183 | enumerators that you have may be invalid (since the node
|
| 184 | may be gone).
|
| 185 | */
|
| 186 | __out_opt POSITION CBaseList::GetHeadPositionI() const
|
| 187 | {
|
| 188 | return (POSITION) m_pFirst;
|
| 189 | } // GetHeadPosition
|
| 190 |
|
| 191 |
|
| 192 |
|
| 193 | __out_opt POSITION CBaseList::GetTailPositionI() const
|
| 194 | {
|
| 195 | return (POSITION) m_pLast;
|
| 196 | } // GetTailPosition
|
| 197 |
|
| 198 |
|
| 199 |
|
| 200 | /* Get the number of objects in the list,
|
| 201 | Get the lock before accessing the count.
|
| 202 | Locking may not be entirely necessary but it has the side effect
|
| 203 | of making sure that all operations are complete before we get it.
|
| 204 | So for example if a list is being added to this list then that
|
| 205 | will have completed in full before we continue rather than seeing
|
| 206 | an intermediate albeit valid state
|
| 207 | */
|
| 208 | int CBaseList::GetCountI() const
|
| 209 | {
|
| 210 | return m_Count;
|
| 211 | } // GetCount
|
| 212 |
|
| 213 |
|
| 214 |
|
| 215 | /* Return the object at rp, update rp to the next object from
|
| 216 | the list or NULL if you have moved over the last object.
|
| 217 | You may still call this function once we return NULL but
|
| 218 | we will continue to return a NULL position value
|
| 219 | */
|
| 220 | __out void *CBaseList::GetNextI(__inout POSITION& rp) const
|
| 221 | {
|
| 222 | /* have we reached the end of the list */
|
| 223 |
|
| 224 | if (rp == NULL) {
|
| 225 | return NULL;
|
| 226 | }
|
| 227 |
|
| 228 | /* Lock the object before continuing */
|
| 229 |
|
| 230 | void *pObject;
|
| 231 |
|
| 232 | /* Copy the original position then step on */
|
| 233 |
|
| 234 | CNode *pn = (CNode *) rp;
|
| 235 | ASSERT(pn != NULL);
|
| 236 | rp = (POSITION) pn->Next();
|
| 237 |
|
| 238 | /* Get the object at the original position from the list */
|
| 239 |
|
| 240 | pObject = pn->GetData();
|
| 241 | // ASSERT(pObject != NULL); // NULL pointers in the list are allowed.
|
| 242 | return pObject;
|
| 243 | } //GetNext
|
| 244 |
|
| 245 |
|
| 246 |
|
| 247 | /* Return the object at p.
|
| 248 | Asking for the object at NULL ASSERTs then returns NULL
|
| 249 | The object is NOT locked. The list is not being changed
|
| 250 | in any way. If another thread is busy deleting the object
|
| 251 | then locking would only result in a change from one bad
|
| 252 | behaviour to another.
|
| 253 | */
|
| 254 | __out_opt void *CBaseList::GetI(__in_opt POSITION p) const
|
| 255 | {
|
| 256 | if (p == NULL) {
|
| 257 | return NULL;
|
| 258 | }
|
| 259 |
|
| 260 | CNode * pn = (CNode *) p;
|
| 261 | void *pObject = pn->GetData();
|
| 262 | // ASSERT(pObject != NULL); // NULL pointers in the list are allowed.
|
| 263 | return pObject;
|
| 264 | } //Get
|
| 265 |
|
| 266 | __out void *CBaseList::GetValidI(__in POSITION p) const
|
| 267 | {
|
| 268 | CNode * pn = (CNode *) p;
|
| 269 | void *pObject = pn->GetData();
|
| 270 | // ASSERT(pObject != NULL); // NULL pointers in the list are allowed.
|
| 271 | return pObject;
|
| 272 | } //Get
|
| 273 |
|
| 274 |
|
| 275 | /* Return the first position in the list which holds the given pointer.
|
| 276 | Return NULL if it's not found.
|
| 277 | */
|
| 278 | __out_opt POSITION CBaseList::FindI( __in void * pObj) const
|
| 279 | {
|
| 280 | POSITION pn;
|
| 281 | INTERNALTRAVERSELIST(*this, pn){
|
| 282 | if (GetI(pn)==pObj) {
|
| 283 | return pn;
|
| 284 | }
|
| 285 | }
|
| 286 | return NULL;
|
| 287 | } // Find
|
| 288 |
|
| 289 |
|
| 290 |
|
| 291 | /* Remove the first node in the list (deletes the pointer to its object
|
| 292 | from the list, does not free the object itself).
|
| 293 | Return the pointer to its object or NULL if empty
|
| 294 | */
|
| 295 | __out_opt void *CBaseList::RemoveHeadI()
|
| 296 | {
|
| 297 | /* All we do is get the head position and ask for that to be deleted.
|
| 298 | We could special case this since some of the code path checking
|
| 299 | in Remove() is redundant as we know there is no previous
|
| 300 | node for example but it seems to gain little over the
|
| 301 | added complexity
|
| 302 | */
|
| 303 |
|
| 304 | return RemoveI((POSITION)m_pFirst);
|
| 305 | } // RemoveHead
|
| 306 |
|
| 307 |
|
| 308 |
|
| 309 | /* Remove the last node in the list (deletes the pointer to its object
|
| 310 | from the list, does not free the object itself).
|
| 311 | Return the pointer to its object or NULL if empty
|
| 312 | */
|
| 313 | __out_opt void *CBaseList::RemoveTailI()
|
| 314 | {
|
| 315 | /* All we do is get the tail position and ask for that to be deleted.
|
| 316 | We could special case this since some of the code path checking
|
| 317 | in Remove() is redundant as we know there is no previous
|
| 318 | node for example but it seems to gain little over the
|
| 319 | added complexity
|
| 320 | */
|
| 321 |
|
| 322 | return RemoveI((POSITION)m_pLast);
|
| 323 | } // RemoveTail
|
| 324 |
|
| 325 |
|
| 326 |
|
| 327 | /* Remove the pointer to the object in this position from the list.
|
| 328 | Deal with all the chain pointers
|
| 329 | Return a pointer to the object removed from the list.
|
| 330 | The node object that is freed as a result
|
| 331 | of this operation is added to the node cache where
|
| 332 | it can be used again.
|
| 333 | Remove(NULL) is a harmless no-op - but probably is a wart.
|
| 334 | */
|
| 335 | __out_opt void *CBaseList::RemoveI(__in_opt POSITION pos)
|
| 336 | {
|
| 337 | /* Lock the critical section before continuing */
|
| 338 |
|
| 339 | // ASSERT (pos!=NULL); // Removing NULL is to be harmless!
|
| 340 | if (pos==NULL) return NULL;
|
| 341 |
|
| 342 |
|
| 343 | CNode *pCurrent = (CNode *) pos;
|
| 344 | ASSERT(pCurrent != NULL);
|
| 345 |
|
| 346 | /* Update the previous node */
|
| 347 |
|
| 348 | CNode *pNode = pCurrent->Prev();
|
| 349 | if (pNode == NULL) {
|
| 350 | m_pFirst = pCurrent->Next();
|
| 351 | } else {
|
| 352 | pNode->SetNext(pCurrent->Next());
|
| 353 | }
|
| 354 |
|
| 355 | /* Update the following node */
|
| 356 |
|
| 357 | pNode = pCurrent->Next();
|
| 358 | if (pNode == NULL) {
|
| 359 | m_pLast = pCurrent->Prev();
|
| 360 | } else {
|
| 361 | pNode->SetPrev(pCurrent->Prev());
|
| 362 | }
|
| 363 |
|
| 364 | /* Get the object this node was looking after */
|
| 365 |
|
| 366 | void *pObject = pCurrent->GetData();
|
| 367 |
|
| 368 | // ASSERT(pObject != NULL); // NULL pointers in the list are allowed.
|
| 369 |
|
| 370 | /* Try and add the node object to the cache -
|
| 371 | a NULL return code from the cache means we ran out of room.
|
| 372 | The cache size is fixed by a constructor argument when the
|
| 373 | list is created and defaults to DEFAULTCACHE.
|
| 374 | This means that the cache will have room for this many
|
| 375 | node objects. So if you have a list of media samples
|
| 376 | and you know there will never be more than five active at
|
| 377 | any given time of them for example then override the default
|
| 378 | constructor
|
| 379 | */
|
| 380 |
|
| 381 | m_Cache.AddToCache(pCurrent);
|
| 382 |
|
| 383 | /* If the list is empty then reset the list event */
|
| 384 |
|
| 385 | --m_Count;
|
| 386 | ASSERT(m_Count >= 0);
|
| 387 | return pObject;
|
| 388 | } // Remove
|
| 389 |
|
| 390 |
|
| 391 |
|
| 392 | /* Add this object to the tail end of our list
|
| 393 | Return the new tail position.
|
| 394 | */
|
| 395 |
|
| 396 | __out_opt POSITION CBaseList::AddTailI(__in void *pObject)
|
| 397 | {
|
| 398 | /* Lock the critical section before continuing */
|
| 399 |
|
| 400 | CNode *pNode;
|
| 401 | // ASSERT(pObject); // NULL pointers in the list are allowed.
|
| 402 |
|
| 403 | /* If there is a node objects in the cache then use
|
| 404 | that otherwise we will have to create a new one */
|
| 405 |
|
| 406 | pNode = (CNode *) m_Cache.RemoveFromCache();
|
| 407 | if (pNode == NULL) {
|
| 408 | pNode = new CNode;
|
| 409 | }
|
| 410 |
|
| 411 | /* Check we have a valid object */
|
| 412 |
|
| 413 | if (pNode == NULL) {
|
| 414 | return NULL;
|
| 415 | }
|
| 416 |
|
| 417 | /* Initialise all the CNode object
|
| 418 | just in case it came from the cache
|
| 419 | */
|
| 420 |
|
| 421 | pNode->SetData(pObject);
|
| 422 | pNode->SetNext(NULL);
|
| 423 | pNode->SetPrev(m_pLast);
|
| 424 |
|
| 425 | if (m_pLast == NULL) {
|
| 426 | m_pFirst = pNode;
|
| 427 | } else {
|
| 428 | m_pLast->SetNext(pNode);
|
| 429 | }
|
| 430 |
|
| 431 | /* Set the new last node pointer and also increment the number
|
| 432 | of list entries, the critical section is unlocked when we
|
| 433 | exit the function
|
| 434 | */
|
| 435 |
|
| 436 | m_pLast = pNode;
|
| 437 | ++m_Count;
|
| 438 |
|
| 439 | return (POSITION) pNode;
|
| 440 | } // AddTail(object)
|
| 441 |
|
| 442 |
|
| 443 |
|
| 444 | /* Add this object to the head end of our list
|
| 445 | Return the new head position.
|
| 446 | */
|
| 447 | __out_opt POSITION CBaseList::AddHeadI(__in void *pObject)
|
| 448 | {
|
| 449 | CNode *pNode;
|
| 450 | // ASSERT(pObject); // NULL pointers in the list are allowed.
|
| 451 |
|
| 452 | /* If there is a node objects in the cache then use
|
| 453 | that otherwise we will have to create a new one */
|
| 454 |
|
| 455 | pNode = (CNode *) m_Cache.RemoveFromCache();
|
| 456 | if (pNode == NULL) {
|
| 457 | pNode = new CNode;
|
| 458 | }
|
| 459 |
|
| 460 | /* Check we have a valid object */
|
| 461 |
|
| 462 | if (pNode == NULL) {
|
| 463 | return NULL;
|
| 464 | }
|
| 465 |
|
| 466 | /* Initialise all the CNode object
|
| 467 | just in case it came from the cache
|
| 468 | */
|
| 469 |
|
| 470 | pNode->SetData(pObject);
|
| 471 |
|
| 472 | /* chain it in (set four pointers) */
|
| 473 | pNode->SetPrev(NULL);
|
| 474 | pNode->SetNext(m_pFirst);
|
| 475 |
|
| 476 | if (m_pFirst == NULL) {
|
| 477 | m_pLast = pNode;
|
| 478 | } else {
|
| 479 | m_pFirst->SetPrev(pNode);
|
| 480 | }
|
| 481 | m_pFirst = pNode;
|
| 482 |
|
| 483 | ++m_Count;
|
| 484 |
|
| 485 | return (POSITION) pNode;
|
| 486 | } // AddHead(object)
|
| 487 |
|
| 488 |
|
| 489 |
|
| 490 | /* Add all the elements in *pList to the tail of this list.
|
| 491 | Return TRUE if it all worked, FALSE if it didn't.
|
| 492 | If it fails some elements may have been added.
|
| 493 | */
|
| 494 | BOOL CBaseList::AddTail(__in CBaseList *pList)
|
| 495 | {
|
| 496 | /* lock the object before starting then enumerate
|
| 497 | each entry in the source list and add them one by one to
|
| 498 | our list (while still holding the object lock)
|
| 499 | Lock the other list too.
|
| 500 | */
|
| 501 | POSITION pos = pList->GetHeadPositionI();
|
| 502 |
|
| 503 | while (pos) {
|
| 504 | if (NULL == AddTailI(pList->GetNextI(pos))) {
|
| 505 | return FALSE;
|
| 506 | }
|
| 507 | }
|
| 508 | return TRUE;
|
| 509 | } // AddTail(list)
|
| 510 |
|
| 511 |
|
| 512 |
|
| 513 | /* Add all the elements in *pList to the head of this list.
|
| 514 | Return TRUE if it all worked, FALSE if it didn't.
|
| 515 | If it fails some elements may have been added.
|
| 516 | */
|
| 517 | BOOL CBaseList::AddHead(__in CBaseList *pList)
|
| 518 | {
|
| 519 | /* lock the object before starting then enumerate
|
| 520 | each entry in the source list and add them one by one to
|
| 521 | our list (while still holding the object lock)
|
| 522 | Lock the other list too.
|
| 523 |
|
| 524 | To avoid reversing the list, traverse it backwards.
|
| 525 | */
|
| 526 |
|
| 527 | POSITION pos;
|
| 528 |
|
| 529 | INTERNALREVERSETRAVERSELIST(*pList, pos) {
|
| 530 | if (NULL== AddHeadI(pList->GetValidI(pos))){
|
| 531 | return FALSE;
|
| 532 | }
|
| 533 | }
|
| 534 | return TRUE;
|
| 535 | } // AddHead(list)
|
| 536 |
|
| 537 |
|
| 538 |
|
| 539 | /* Add the object after position p
|
| 540 | p is still valid after the operation.
|
| 541 | AddAfter(NULL,x) adds x to the start - same as AddHead
|
| 542 | Return the position of the new object, NULL if it failed
|
| 543 | */
|
| 544 | __out_opt POSITION CBaseList::AddAfterI(__in_opt POSITION pos, __in void * pObj)
|
| 545 | {
|
| 546 | if (pos==NULL)
|
| 547 | return AddHeadI(pObj);
|
| 548 |
|
| 549 | /* As someone else might be furkling with the list -
|
| 550 | Lock the critical section before continuing
|
| 551 | */
|
| 552 | CNode *pAfter = (CNode *) pos;
|
| 553 | ASSERT(pAfter != NULL);
|
| 554 | if (pAfter==m_pLast)
|
| 555 | return AddTailI(pObj);
|
| 556 |
|
| 557 | /* set pnode to point to a new node, preferably from the cache */
|
| 558 |
|
| 559 | CNode *pNode = (CNode *) m_Cache.RemoveFromCache();
|
| 560 | if (pNode == NULL) {
|
| 561 | pNode = new CNode;
|
| 562 | }
|
| 563 |
|
| 564 | /* Check we have a valid object */
|
| 565 |
|
| 566 | if (pNode == NULL) {
|
| 567 | return NULL;
|
| 568 | }
|
| 569 |
|
| 570 | /* Initialise all the CNode object
|
| 571 | just in case it came from the cache
|
| 572 | */
|
| 573 |
|
| 574 | pNode->SetData(pObj);
|
| 575 |
|
| 576 | /* It is to be added to the middle of the list - there is a before
|
| 577 | and after node. Chain it after pAfter, before pBefore.
|
| 578 | */
|
| 579 | CNode * pBefore = pAfter->Next();
|
| 580 | ASSERT(pBefore != NULL);
|
| 581 |
|
| 582 | /* chain it in (set four pointers) */
|
| 583 | pNode->SetPrev(pAfter);
|
| 584 | pNode->SetNext(pBefore);
|
| 585 | pBefore->SetPrev(pNode);
|
| 586 | pAfter->SetNext(pNode);
|
| 587 |
|
| 588 | ++m_Count;
|
| 589 |
|
| 590 | return (POSITION) pNode;
|
| 591 |
|
| 592 | } // AddAfter(object)
|
| 593 |
|
| 594 |
|
| 595 |
|
| 596 | BOOL CBaseList::AddAfter(__in_opt POSITION p, __in CBaseList *pList)
|
| 597 | {
|
| 598 | POSITION pos;
|
| 599 | INTERNALTRAVERSELIST(*pList, pos) {
|
| 600 | /* p follows along the elements being added */
|
| 601 | p = AddAfterI(p, pList->GetValidI(pos));
|
| 602 | if (p==NULL) return FALSE;
|
| 603 | }
|
| 604 | return TRUE;
|
| 605 | } // AddAfter(list)
|
| 606 |
|
| 607 |
|
| 608 |
|
| 609 | /* Mirror images:
|
| 610 | Add the element or list after position p.
|
| 611 | p is still valid after the operation.
|
| 612 | AddBefore(NULL,x) adds x to the end - same as AddTail
|
| 613 | */
|
| 614 | __out_opt POSITION CBaseList::AddBeforeI(__in_opt POSITION pos, __in void * pObj)
|
| 615 | {
|
| 616 | if (pos==NULL)
|
| 617 | return AddTailI(pObj);
|
| 618 |
|
| 619 | /* set pnode to point to a new node, preferably from the cache */
|
| 620 |
|
| 621 | CNode *pBefore = (CNode *) pos;
|
| 622 | ASSERT(pBefore != NULL);
|
| 623 | if (pBefore==m_pFirst)
|
| 624 | return AddHeadI(pObj);
|
| 625 |
|
| 626 | CNode * pNode = (CNode *) m_Cache.RemoveFromCache();
|
| 627 | if (pNode == NULL) {
|
| 628 | pNode = new CNode;
|
| 629 | }
|
| 630 |
|
| 631 | /* Check we have a valid object */
|
| 632 |
|
| 633 | if (pNode == NULL) {
|
| 634 | return NULL;
|
| 635 | }
|
| 636 |
|
| 637 | /* Initialise all the CNode object
|
| 638 | just in case it came from the cache
|
| 639 | */
|
| 640 |
|
| 641 | pNode->SetData(pObj);
|
| 642 |
|
| 643 | /* It is to be added to the middle of the list - there is a before
|
| 644 | and after node. Chain it after pAfter, before pBefore.
|
| 645 | */
|
| 646 |
|
| 647 | CNode * pAfter = pBefore->Prev();
|
| 648 | ASSERT(pAfter != NULL);
|
| 649 |
|
| 650 | /* chain it in (set four pointers) */
|
| 651 | pNode->SetPrev(pAfter);
|
| 652 | pNode->SetNext(pBefore);
|
| 653 | pBefore->SetPrev(pNode);
|
| 654 | pAfter->SetNext(pNode);
|
| 655 |
|
| 656 | ++m_Count;
|
| 657 |
|
| 658 | return (POSITION) pNode;
|
| 659 |
|
| 660 | } // Addbefore(object)
|
| 661 |
|
| 662 |
|
| 663 |
|
| 664 | BOOL CBaseList::AddBefore(__in_opt POSITION p, __in CBaseList *pList)
|
| 665 | {
|
| 666 | POSITION pos;
|
| 667 | INTERNALREVERSETRAVERSELIST(*pList, pos) {
|
| 668 | /* p follows along the elements being added */
|
| 669 | p = AddBeforeI(p, pList->GetValidI(pos));
|
| 670 | if (p==NULL) return FALSE;
|
| 671 | }
|
| 672 | return TRUE;
|
| 673 | } // AddBefore(list)
|
| 674 |
|
| 675 |
|
| 676 |
|
| 677 | /* Split *this after position p in *this
|
| 678 | Retain as *this the tail portion of the original *this
|
| 679 | Add the head portion to the tail end of *pList
|
| 680 | Return TRUE if it all worked, FALSE if it didn't.
|
| 681 |
|
| 682 | e.g.
|
| 683 | foo->MoveToTail(foo->GetHeadPosition(), bar);
|
| 684 | moves one element from the head of foo to the tail of bar
|
| 685 | foo->MoveToTail(NULL, bar);
|
| 686 | is a no-op
|
| 687 | foo->MoveToTail(foo->GetTailPosition, bar);
|
| 688 | concatenates foo onto the end of bar and empties foo.
|
| 689 |
|
| 690 | A better, except excessively long name might be
|
| 691 | MoveElementsFromHeadThroughPositionToOtherTail
|
| 692 | */
|
| 693 | BOOL CBaseList::MoveToTail
|
| 694 | (__in_opt POSITION pos, __in CBaseList *pList)
|
| 695 | {
|
| 696 | /* Algorithm:
|
| 697 | Note that the elements (including their order) in the concatenation
|
| 698 | of *pList to the head of *this is invariant.
|
| 699 | 1. Count elements to be moved
|
| 700 | 2. Join *pList onto the head of this to make one long chain
|
| 701 | 3. Set first/Last pointers in *this and *pList
|
| 702 | 4. Break the chain at the new place
|
| 703 | 5. Adjust counts
|
| 704 | 6. Set/Reset any events
|
| 705 | */
|
| 706 |
|
| 707 | if (pos==NULL) return TRUE; // no-op. Eliminates special cases later.
|
| 708 |
|
| 709 |
|
| 710 | /* Make cMove the number of nodes to move */
|
| 711 | CNode * p = (CNode *)pos;
|
| 712 | int cMove = 0; // number of nodes to move
|
| 713 | while(p!=NULL) {
|
| 714 | p = p->Prev();
|
| 715 | ++cMove;
|
| 716 | }
|
| 717 |
|
| 718 |
|
| 719 | /* Join the two chains together */
|
| 720 | if (pList->m_pLast!=NULL)
|
| 721 | pList->m_pLast->SetNext(m_pFirst);
|
| 722 | if (m_pFirst!=NULL)
|
| 723 | m_pFirst->SetPrev(pList->m_pLast);
|
| 724 |
|
| 725 |
|
| 726 | /* set first and last pointers */
|
| 727 | p = (CNode *)pos;
|
| 728 |
|
| 729 | if (pList->m_pFirst==NULL)
|
| 730 | pList->m_pFirst = m_pFirst;
|
| 731 | m_pFirst = p->Next();
|
| 732 | if (m_pFirst==NULL)
|
| 733 | m_pLast = NULL;
|
| 734 | pList->m_pLast = p;
|
| 735 |
|
| 736 |
|
| 737 | /* Break the chain after p to create the new pieces */
|
| 738 | if (m_pFirst!=NULL)
|
| 739 | m_pFirst->SetPrev(NULL);
|
| 740 | p->SetNext(NULL);
|
| 741 |
|
| 742 |
|
| 743 | /* Adjust the counts */
|
| 744 | m_Count -= cMove;
|
| 745 | pList->m_Count += cMove;
|
| 746 |
|
| 747 | return TRUE;
|
| 748 |
|
| 749 | } // MoveToTail
|
| 750 |
|
| 751 |
|
| 752 |
|
| 753 | /* Mirror image of MoveToTail:
|
| 754 | Split *this before position p in *this.
|
| 755 | Retain in *this the head portion of the original *this
|
| 756 | Add the tail portion to the start (i.e. head) of *pList
|
| 757 | Return TRUE if it all worked, FALSE if it didn't.
|
| 758 |
|
| 759 | e.g.
|
| 760 | foo->MoveToHead(foo->GetTailPosition(), bar);
|
| 761 | moves one element from the tail of foo to the head of bar
|
| 762 | foo->MoveToHead(NULL, bar);
|
| 763 | is a no-op
|
| 764 | foo->MoveToHead(foo->GetHeadPosition, bar);
|
| 765 | concatenates foo onto the start of bar and empties foo.
|
| 766 | */
|
| 767 | BOOL CBaseList::MoveToHead
|
| 768 | (__in_opt POSITION pos, __in CBaseList *pList)
|
| 769 | {
|
| 770 |
|
| 771 | /* See the comments on the algorithm in MoveToTail */
|
| 772 |
|
| 773 | if (pos==NULL) return TRUE; // no-op. Eliminates special cases later.
|
| 774 |
|
| 775 | /* Make cMove the number of nodes to move */
|
| 776 | CNode * p = (CNode *)pos;
|
| 777 | int cMove = 0; // number of nodes to move
|
| 778 | while(p!=NULL) {
|
| 779 | p = p->Next();
|
| 780 | ++cMove;
|
| 781 | }
|
| 782 |
|
| 783 |
|
| 784 | /* Join the two chains together */
|
| 785 | if (pList->m_pFirst!=NULL)
|
| 786 | pList->m_pFirst->SetPrev(m_pLast);
|
| 787 | if (m_pLast!=NULL)
|
| 788 | m_pLast->SetNext(pList->m_pFirst);
|
| 789 |
|
| 790 |
|
| 791 | /* set first and last pointers */
|
| 792 | p = (CNode *)pos;
|
| 793 |
|
| 794 |
|
| 795 | if (pList->m_pLast==NULL)
|
| 796 | pList->m_pLast = m_pLast;
|
| 797 |
|
| 798 | m_pLast = p->Prev();
|
| 799 | if (m_pLast==NULL)
|
| 800 | m_pFirst = NULL;
|
| 801 | pList->m_pFirst = p;
|
| 802 |
|
| 803 |
|
| 804 | /* Break the chain after p to create the new pieces */
|
| 805 | if (m_pLast!=NULL)
|
| 806 | m_pLast->SetNext(NULL);
|
| 807 | p->SetPrev(NULL);
|
| 808 |
|
| 809 |
|
| 810 | /* Adjust the counts */
|
| 811 | m_Count -= cMove;
|
| 812 | pList->m_Count += cMove;
|
| 813 |
|
| 814 | return TRUE;
|
| 815 |
|
| 816 | } // MoveToHead
|
| 817 |
|
| 818 |
|
| 819 |
|
| 820 | /* Reverse the order of the [pointers to] objects in *this
|
| 821 | */
|
| 822 | void CBaseList::Reverse()
|
| 823 | {
|
| 824 | /* algorithm:
|
| 825 | The obvious booby trap is that you flip pointers around and lose
|
| 826 | addressability to the node that you are going to process next.
|
| 827 | The easy way to avoid this is do do one chain at a time.
|
| 828 |
|
| 829 | Run along the forward chain,
|
| 830 | For each node, set the reverse pointer to the one ahead of us.
|
| 831 | The reverse chain is now a copy of the old forward chain, including
|
| 832 | the NULL termination.
|
| 833 |
|
| 834 | Run along the reverse chain (i.e. old forward chain again)
|
| 835 | For each node set the forward pointer of the node ahead to point back
|
| 836 | to the one we're standing on.
|
| 837 | The first node needs special treatment,
|
| 838 | it's new forward pointer is NULL.
|
| 839 | Finally set the First/Last pointers
|
| 840 |
|
| 841 | */
|
| 842 | CNode * p;
|
| 843 |
|
| 844 | // Yes we COULD use a traverse, but it would look funny!
|
| 845 | p = m_pFirst;
|
| 846 | while (p!=NULL) {
|
| 847 | CNode * q;
|
| 848 | q = p->Next();
|
| 849 | p->SetNext(p->Prev());
|
| 850 | p->SetPrev(q);
|
| 851 | p = q;
|
| 852 | }
|
| 853 |
|
| 854 | p = m_pFirst;
|
| 855 | m_pFirst = m_pLast;
|
| 856 | m_pLast = p;
|
| 857 |
|
| 858 |
|
| 859 | #if 0 // old version
|
| 860 |
|
| 861 | if (m_pFirst==NULL) return; // empty list
|
| 862 | if (m_pFirst->Next()==NULL) return; // single node list
|
| 863 |
|
| 864 |
|
| 865 | /* run along forward chain */
|
| 866 | for ( p = m_pFirst
|
| 867 | ; p!=NULL
|
| 868 | ; p = p->Next()
|
| 869 | ){
|
| 870 | p->SetPrev(p->Next());
|
| 871 | }
|
| 872 |
|
| 873 |
|
| 874 | /* special case first element */
|
| 875 | m_pFirst->SetNext(NULL); // fix the old first element
|
| 876 |
|
| 877 |
|
| 878 | /* run along new reverse chain i.e. old forward chain again */
|
| 879 | for ( p = m_pFirst // start at the old first element
|
| 880 | ; p->Prev()!=NULL // while there's a node still to be set
|
| 881 | ; p = p->Prev() // work in the same direction as before
|
| 882 | ){
|
| 883 | p->Prev()->SetNext(p);
|
| 884 | }
|
| 885 |
|
| 886 |
|
| 887 | /* fix forward and reverse pointers
|
| 888 | - the triple XOR swap would work but all the casts look hideous */
|
| 889 | p = m_pFirst;
|
| 890 | m_pFirst = m_pLast;
|
| 891 | m_pLast = p;
|
| 892 | #endif
|
| 893 |
|
| 894 | } // Reverse
|
| 895 |
|
| 896 | #endif /* PJMEDIA_VIDEO_DEV_HAS_DSHOW */
|