蔡军生先生第二人生的源码分析(110)脚本的语法分析（5）

末蓝、 2022-08-27 12:54 130阅读 0赞

在脚本里有很多语句组成的，那么对于语句又是怎么进行语法分析的呢？下面就来分析for语句的分析过程，它的声明代码如下：

\#001  class LLScriptFor : public LLScriptStatement

\#002  \{

\#003  public:

构造函数，输入保存的值。

\#004     LLScriptFor(S32 line, S32 col, LLScriptExpression \*sequence, LLScriptExpression \*expression, LLScriptExpression \*expressionlist, LLScriptStatement \*statement)

\#005         : LLScriptStatement(line, col, LSSMT\_FOR), mSequence(sequence), mExpression(expression), mExpressionList(expressionlist), mStatement(statement),

\#006  mType(LST\_NULL)

\#007     \{

\#008     \}

\#009

析构造函数。

\#010     ~LLScriptFor()

\#011     \{

\#012     \}

\#013

递归处理函数。

\#014     void recurse(FILE \*fp, S32 tabs, S32 tabsize, LSCRIPTCompilePass pass, LSCRIPTPruneType ptype, BOOL &prunearg, LLScriptScope \*scope, LSCRIPTType &type,

\#015  LSCRIPTType basetype, U64 &count, LLScriptByteCodeChunk \*chunk, LLScriptByteCodeChunk \*heap, S32 stacksize, LLScriptScopeEntry \*entry, S32 entrycount, LLScriptLibData \*\*ldata);

获取语句生成代码的大小。

\#016     S32 getSize();

\#017

初始化表达式。

\#018     LLScriptExpression      \*mSequence;

条件判断表达式。

\#019     LLScriptExpression      \*mExpression;

最后执行的表达式列表。

\#020     LLScriptExpression      \*mExpressionList;

循环体的语句块。

\#021     LLScriptStatement       \*mStatement;

脚本的类型。

\#022     LSCRIPTType             mType;

\#023  \};

下面来分析它的递归调用函数的处理：

\#001  void LLScriptFor::recurse(FILE \*fp, S32 tabs, S32 tabsize, LSCRIPTCompilePass pass, LSCRIPTPruneType ptype, BOOL &prunearg, LLScriptScope \*scope, LSCRIPTType &type,

\#002  LSCRIPTType basetype, U64 &count, LLScriptByteCodeChunk \*chunk, LLScriptByteCodeChunk \*heap, S32 stacksize, LLScriptScopeEntry \*entry, S32 entrycount, LLScriptLibData \*\*ldata)

\#003  \{

如果分析出错就返回。

\#004     if (gErrorToText.getErrors())

\#005     \{

\#006         return;

\#007     \}

下面根据不同的遍来处理。

\#008     switch(pass)

\#009     \{

打印输出脚本的代码。

\#010     case LSCP\_PRETTY\_PRINT:

\#011         fdotabs(fp, tabs, tabsize);

\#012         fprintf(fp, "for ( ");

\#013         if(mSequence)

\#014             mSequence->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#015  NULL);

\#016         fprintf(fp, " ; ");

\#017         mExpression->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount, NULL);

\#018         fprintf(fp, " ; ");

\#019         if(mExpressionList)

\#020             mExpressionList->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry,

\#021  entrycount, NULL);

\#022         fprintf(fp, " )/n");

\#023         if(mStatement)

\#024             mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#025  NULL);

\#026         break;

下面输出编译出来的汇编代码。

\#027     case LSCP\_EMIT\_ASSEMBLY:

\#028         \{

\#029             S32 tjump1 =  gTempJumpCount++;

\#030             S32 tjump2 =  gTempJumpCount++;

\#031             if(mSequence)

\#032                 mSequence->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#033  entry, entrycount, NULL);

\#034             fprintf(fp, "LABEL \#\#Temp Jump %d\#\#/n", tjump1);

\#035             mExpression->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#036  NULL);

\#037             fprintf(fp, "JUMPNIF \#\#Temp Jump %d\#\#/n", tjump2);

\#038             if(mStatement)

\#039                 mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#040  entry, entrycount, NULL);

\#041             if(mExpressionList)

\#042                 mExpressionList->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#043  entry, entrycount, NULL);

\#044             fprintf(fp, "JUMP \#\#Temp Jump %d\#\#/n", tjump1);

\#045             fprintf(fp, "LABEL \#\#Temp Jump %d\#\#/n", tjump2);

\#046         \}

\#047         break;

下面优化代码，如果代码不用运行，就可以删除。

\#048     case LSCP\_PRUNE:

\#049         if (ptype == LSPRUNE\_DEAD\_CODE)

\#050             prunearg = TRUE;

\#051         else

\#052             prunearg = FALSE;

\#053         break;

下面进行for语句的类型检查。

\#054     case LSCP\_TYPE:

\#055         if(mSequence)

\#056             mSequence->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#057  NULL);

\#058         mExpression->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount, NULL);

\#059         mType = type;

\#060         if(mExpressionList)

\#061             mExpressionList->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry,

\#062  entrycount, NULL);

\#063         if(mStatement)

\#064             mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#065  NULL);

\#066         break;

下面生成for语句的目标代码。

\#067     case LSCP\_EMIT\_BYTE\_CODE:

\#068         \{

\#069             char jumpname1\[32\];        /\*Flawfinder: ignore\*/

\#070             snprintf(jumpname1, sizeof(jumpname1), "\#\#Temp Jump %d\#\#", gTempJumpCount++);   /\* Flawfinder: ignore \*/

\#071             char jumpname2\[32\];            /\*Flawfinder: ignore\*/

\#072             snprintf(jumpname2, sizeof(jumpname2), "\#\#Temp Jump %d\#\#", gTempJumpCount++);   /\* Flawfinder: ignore \*/

\#073

\#074             if(mSequence)

\#075                 mSequence->recurse(fp, tabs, tabsize, LSCP\_TO\_STACK, ptype, prunearg, scope, type, basetype, count, chunk, heap,

\#076  stacksize, entry, entrycount, NULL);

\#077             chunk->addLabel(jumpname1);

\#078             mExpression->recurse(fp, tabs, tabsize, LSCP\_TO\_STACK, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry,

\#079  entrycount, NULL);

\#080             chunk->addByte(LSCRIPTOpCodes\[LOPC\_JUMPNIF\]);

\#081             chunk->addByte(LSCRIPTTypeByte\[mType\]);

\#082             chunk->addBytes(LSCRIPTDataSize\[LST\_INTEGER\]);

\#083             chunk->addJump(jumpname2);

\#084             if(mStatement)

\#085                 mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#086  entry, entrycount, NULL);

\#087             if(mExpressionList)

\#088                 mExpressionList->recurse(fp, tabs, tabsize, LSCP\_TO\_STACK, ptype, prunearg, scope, type, basetype, count, chunk,

\#089  heap, stacksize, entry, entrycount, NULL);

\#090             chunk->addByte(LSCRIPTOpCodes\[LOPC\_JUMP\]);

\#091             chunk->addBytes(LSCRIPTDataSize\[LST\_INTEGER\]);

\#092             chunk->addJump(jumpname1);

\#093             chunk->addLabel(jumpname2);

\#094         \}

\#095         break;

下面生成CIL的汇编代码。

\#096     case LSCP\_EMIT\_CIL\_ASSEMBLY:

\#097         \{

\#098             S32 tjump1 =  gTempJumpCount++;

\#099             S32 tjump2 =  gTempJumpCount++;

\#100             if(mSequence)

\#101                 mSequence->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#102  entry, entrycount, NULL);

\#103             fprintf(fp, "LabelTempJump%d:/n", tjump1);

\#104             mExpression->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#105  NULL);

\#106             fprintf(fp, "brfalse LabelTempJump%d/n", tjump2);

\#107             if(mStatement)

\#108                 mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#109  entry, entrycount, NULL);

\#110             if(mExpressionList)

\#111                 mExpressionList->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize,

\#112  entry, entrycount, NULL);

\#113             fprintf(fp, "br LabelTempJump%d/n", tjump1);

\#114             fprintf(fp, "LabelTempJump%d:/n", tjump2);

\#115         \}

\#116         break;

下面进行递归处理for里出现的表达式。

\#117     default:

\#118         if(mSequence)

\#119             mSequence->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#120  NULL);

\#121         mExpression->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount, NULL);

\#122         if(mExpressionList)

\#123             mExpressionList->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry,

\#124  entrycount, NULL);

\#125         if(mStatement)

\#126             mStatement->recurse(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount,

\#127  NULL);

\#128         break;

\#129     \}

\#130     gonext(fp, tabs, tabsize, pass, ptype, prunearg, scope, type, basetype, count, chunk, heap, stacksize, entry, entrycount, NULL);

\#131  \}

在语法分析里，把所有合法的语句都生成一个合适的类对象，然后通过遍历所有类对象进行类型检查，或者生成代码。所有其它语句都是这样进行的，比如语句if就是使用类LLScriptIf来保存和分析的。当检查所有脚本都是合法的语法后，就生成了语法树，后面通过遍历语法树来生成代码。