libmxml數據結構（源碼分析）
阿新 • • 發佈：2019-04-05
attrs def device same next 下層 election str 一個
libmxml是一個開源、小巧的C語言xml庫。這裏簡單分析一下它是用什麽樣的數據結構來保存分析過的xml文檔。
mxml關鍵的結構體mxml_node_t是這樣的實現的：
struct mxml_node_s /**** An XML node. @[email protected] ****/ { mxml_type_t type; /* Node type */ struct mxml_node_s *next; /* Next node under same parent */ structmxml_node_s *prev; /* Previous node under same parent */ struct mxml_node_s *parent; /* Parent node */ struct mxml_node_s *child; /* First child node */ struct mxml_node_s *last_child; /* Last child node */ mxml_value_t value; /* Node value */ int ref_count; /*Use count */ void *user_data; /* User data */ }; typedef struct mxml_node_s mxml_node_t; /**** An XML node. ****/
它使用左孩子右兄弟的樹形結構來描述xml報文：即下層節點登記在child鏈表，兄弟節點登記在next鏈表。 如果某個節點下面有N個子節點，則child指向第一個子節點，該子節點的next指向下一個同父節點的子節點。 比較特殊的是，mxml把xml節點值也認為是一個子節點。例如<group>value</group>， 其中value(type是MXML_OPAQUE)是一個獨立的子節點，掛載在group節點(type是MXML_ELEMENT)下面。 另外，空白符（空格，回車換行，制表符）和註釋，雖然對xml報文無實質意義，但mxml還是把它們做為一個節點存儲起來。
由於mxml只是使用簡單的鏈表存儲xml元素，所以元素節點個數比較多時，mxml查找元素效率是比較低的。所以libmxml提供了一個索引查找的函數，它需要先遍歷xml元素樹，生成一個排序過的數組，加快查找速度。
為了方便大家理解，我寫了一個函數打印xml結構體。
void printNode(mxml_node_t *node, int nNodeSn, int level) { static int currNodeSn = 0; if (node == NULL) { return; } ++currNodeSn; //每遇到一個新節點 則將節點序號遞增，做為本節點序號 printf("[%- 3d -> %- 3d] ", currNodeSn, nNodeSn); switch (node->type) { case MXML_ELEMENT: { int i; printf("level %d MXML_ELEMENT [%s]", level, node->value.element.name); for (i = 0; i < node->value.element.num_attrs; ++i) { printf(" %s=%s", node->value.element.attrs[i].name, node->value.element.attrs[i].value); } printf("\n"); } break; case MXML_INTEGER: printf("level %d MXML_INTEGER %d\n", level, node->value.integer); break; case MXML_OPAQUE: printf("level %d MXML_OPAQUE [%s]\n", level, node->value.opaque); break; case MXML_REAL: printf("level %d MXML_REAL %lf\n", level, node->value.real); break; case MXML_TEXT: printf("level %d MXML_TEXT [%s]\n", level, node->value.text.string); break; case MXML_CUSTOM: printf("level %d MXML_CUSTOM\n", level); break; default: printf("unknown node type %d\n", node->type); } //深度優先遍歷 if (node->child) { //訪問子節點時把本節點序號做為父節點序號 層級加1 printNode(node->child, currNodeSn, level + 1); } if (node->next) { //訪問兄弟節點，直接傳父節點序號即可 層級也不用加1 printNode(node->next, nNodeSn, level); } }
運行示例如下：
xml源如下：
<?xml version="1.0" encoding="GBK" ?> <group> <option>122334 我們 <string>我們</string>45677 <keyword type="opaque">InputSlot</keyword> <default type="opaque">Auto</default> <text>Media Source</text> <order type="real">10.000000</order> <choice> <keyword type="opaque">Auto</keyword> <text>Auto Tray Selection</text> <code type="opaque" /> </choice> <choice> <keyword type="opaque">Upper</keyword> <text>Tray 1</text> <code type="opaque"><</MediaPosition 0>>setpagedevice</code> </choice> <choice> <keyword type="opaque">Lower</keyword> <text>Tray 2</text> <code type="opaque"><</MediaPosition 1>>setpagedevice</code> </choice> </option> 我 12334545 050504550 <integer>123</integer> <string>Now is the time for all good men to come to the aid of their country.</string> <!-- this is a comment --> <![CDATA[this is CDATA 0123456789ABCDEF]]> </group>
用我這個printNode分析結果如下：
說明：[ 1 -> 0 ]，代表本節點序號是1，其父節點序號是0，level 0代表本節點是最頂層節點。 [ 1 -> 0 ] level 0 MXML_ELEMENT [?xml version="1.0" encoding="GBK" ?] [ 2 -> 1 ] level 1 MXML_OPAQUE [ ] [ 3 -> 1 ] level 1 MXML_ELEMENT [group] [ 4 -> 3 ] level 2 MXML_OPAQUE [ ] [ 5 -> 3 ] level 2 MXML_ELEMENT [option] [ 6 -> 5 ] level 3 MXML_OPAQUE [122334 我們 ] [ 7 -> 5 ] level 3 MXML_ELEMENT [string] [ 8 -> 7 ] level 4 MXML_OPAQUE [我們] [ 9 -> 5 ] level 3 MXML_OPAQUE [45677 ] [ 10 -> 5 ] level 3 MXML_ELEMENT [keyword] type=opaque [ 11 -> 10] level 4 MXML_OPAQUE [InputSlot] [ 12 -> 5 ] level 3 MXML_OPAQUE [ ] [ 13 -> 5 ] level 3 MXML_ELEMENT [default] type=opaque [ 14 -> 13] level 4 MXML_OPAQUE [Auto] [ 15 -> 5 ] level 3 MXML_OPAQUE [ ] [ 16 -> 5 ] level 3 MXML_ELEMENT [text] [ 17 -> 16] level 4 MXML_OPAQUE [Media Source] [ 18 -> 5 ] level 3 MXML_OPAQUE [ ] [ 19 -> 5 ] level 3 MXML_ELEMENT [order] type=real [ 20 -> 19] level 4 MXML_OPAQUE [10.000000] [ 21 -> 5 ] level 3 MXML_OPAQUE [ ] [ 22 -> 5 ] level 3 MXML_ELEMENT [choice] [ 23 -> 22] level 4 MXML_OPAQUE [ ] [ 24 -> 22] level 4 MXML_ELEMENT [keyword] type=opaque [ 25 -> 24] level 5 MXML_OPAQUE [Auto] [ 26 -> 22] level 4 MXML_OPAQUE [ ] [ 27 -> 22] level 4 MXML_ELEMENT [text] [ 28 -> 27] level 5 MXML_OPAQUE [Auto Tray Selection] [ 29 -> 22] level 4 MXML_OPAQUE [ ] [ 30 -> 22] level 4 MXML_ELEMENT [code] type=opaque [ 31 -> 22] level 4 MXML_OPAQUE [ ] [ 32 -> 5 ] level 3 MXML_OPAQUE [ ] [ 33 -> 5 ] level 3 MXML_ELEMENT [choice] [ 34 -> 33] level 4 MXML_OPAQUE [ ] [ 35 -> 33] level 4 MXML_ELEMENT [keyword] type=opaque [ 36 -> 35] level 5 MXML_OPAQUE [Upper] [ 37 -> 33] level 4 MXML_OPAQUE [ ] [ 38 -> 33] level 4 MXML_ELEMENT [text] [ 39 -> 38] level 5 MXML_OPAQUE [Tray 1] [ 40 -> 33] level 4 MXML_OPAQUE [ ] [ 41 -> 33] level 4 MXML_ELEMENT [code] type=opaque [ 42 -> 41] level 5 MXML_OPAQUE [<</MediaPosition 0>>setpagedevice] [ 43 -> 33] level 4 MXML_OPAQUE [ ] [ 44 -> 5 ] level 3 MXML_OPAQUE [ ] [ 45 -> 5 ] level 3 MXML_ELEMENT [choice] [ 46 -> 45] level 4 MXML_OPAQUE [ ] [ 47 -> 45] level 4 MXML_ELEMENT [keyword] type=opaque [ 48 -> 47] level 5 MXML_OPAQUE [Lower] [ 49 -> 45] level 4 MXML_OPAQUE [ ] [ 50 -> 45] level 4 MXML_ELEMENT [text] [ 51 -> 50] level 5 MXML_OPAQUE [Tray 2] [ 52 -> 45] level 4 MXML_OPAQUE [ ] [ 53 -> 45] level 4 MXML_ELEMENT [code] type=opaque [ 54 -> 53] level 5 MXML_OPAQUE [<</MediaPosition 1>>setpagedevice] [ 55 -> 45] level 4 MXML_OPAQUE [ ] [ 56 -> 5 ] level 3 MXML_OPAQUE [ ] [ 57 -> 3 ] level 2 MXML_OPAQUE [ 我12334545 050504550 ] [ 58 -> 3 ] level 2 MXML_ELEMENT [integer] [ 59 -> 58] level 3 MXML_OPAQUE [123] [ 60 -> 3 ] level 2 MXML_OPAQUE [ ] [ 61 -> 3 ] level 2 MXML_ELEMENT [string] [ 62 -> 61] level 3 MXML_OPAQUE [Now is the time for all good men to come to the aid of their country.] [ 63 -> 3 ] level 2 MXML_OPAQUE [ ] [ 64 -> 3 ] level 2 MXML_ELEMENT [!-- this is a comment --] [ 65 -> 3 ] level 2 MXML_OPAQUE [ ] [ 66 -> 3 ] level 2 MXML_ELEMENT [![CDATA[this is CDATA 0123456789ABCDEF]]] [ 67 -> 3 ] level 2 MXML_OPAQUE [ ]
libmxml數據結構（源碼分析）