Code Too Large For Try Statement

Right_Then

unread,

Jul 21, 2014, 6:23:10 AM7/21/14

to

Dear Fellows,

My grammar was parsing too slow. SLL never worked for me.

Diagnostics showed that it is going for full context search and encountering ambiguities therefore SLL always fails.

line 2:3 reportAttemptingFullContext d=821 (argValuesInFunCall), input='"clang/AST/ASTConsumer.h",'

line 2:3 reportAmbiguity d=821 (argValuesInFunCall): ambigAlts={1, 3}, input='"clang/AST/ASTConsumer.h",'

It showed maximum problem in expression rule.

Because above stated rules and other rules would lead to expression rule also.

Because almost! everything is expression in my language(attempt).

expression rule has 60 options/possibilities and 42 options are in the primary rule of expression

i.e

expression:

primary

| expression + expression

| etc...

;

I dont know about Left factoring as much i do about Refactoring the code.

and that is how i am addressing my grammar so that it remains readable/understandable to me.

I have no background in parsing or compiler design i am just a hobbyist.

So i decided to make things easier for Antlr by coding all the possibilities for expression rule/options myself

and putting them there explicitly. Hoping speed would increase then, because Antlr wont have to spend time guessing expressions.

It turned out the possibilities for those 60 expression options are 125606 if expanded to be exact and counting!!!.

Well i did it! i wrote them all. But i have only placed first 400 expression rule options in the .g4 file. And

javac complains about the above mentioned error as.

java:57461: error: code too large for try statement catch (RecognitionException re) {

^

2. java:41028: error: code too large for try statement

try {

^

3. java:41020: error: code too large

private ExpressionContext expression(int _p) throws RecognitionException {

^

I read Mr. Terrance suggestion in a related question https://groups.google.com/forum/#!topic/antlr-discussion/AkbmtmDtQe0

to divide such rules with many options into different rules so that not to clutter one rule with many x|y|z options.

Alright that can be done but now i am fearing that java has a limit for the size of function itself i.e

A single method in a Java class may be at most 64KB of bytecode.

I tried it now errors are of memory heap

Exception in thread "main" java.lang.reflect.InvocationTargetException

at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

at sun.reflect.NativeMethodAccessorImpl.invoke(Unknown Source)

at sun.reflect.DelegatingMethodAccessorImpl.invoke(Unknown Source)

at java.lang.reflect.Method.invoke(Unknown Source)

at org.antlr.v4.runtime.misc.TestRig.process(TestRig.java:249)

at org.antlr.v4.runtime.misc.TestRig.process(TestRig.java:192)

at org.antlr.v4.runtime.misc.TestRig.main(TestRig.java:143)

Caused by: java.lang.OutOfMemoryError: Java heap space

at org.antlr.v4.runtime.misc.DoubleKeyMap.put(DoubleKeyMap.java:49)

at org.antlr.v4.runtime.atn.PredictionContext.mergeSingletons(PredictionContext.java:259)

at org.antlr.v4.runtime.atn.PredictionContext.merge(PredictionContext.java:180)

at org.antlr.v4.runtime.atn.PredictionContext.mergeArrays(PredictionContext.java:420)

at org.antlr.v4.runtime.atn.PredictionContext.merge(PredictionContext.java:199)

at org.antlr.v4.runtime.atn.ATNConfigSet.add(ATNConfigSet.java:175)

at org.antlr.v4.runtime.atn.ParserATNSimulator.closure_(ParserATNSimulator.java:1447)