/** This code is factored out from mismatched token and mismatched set
* recovery. It handles "single token insertion" error recovery for
* both. No tokens are consumed to recover from insertions. Return
* true if recovery was possible else return false.
*/
static ANTLR3_BOOLEAN
recoverFromMismatchedElement (pANTLR3_BASE_RECOGNIZER recognizer, pANTLR3_BITSET follow)
{
pANTLR3_BITSET viableToksFollowingRule;
pANTLR3_BITSET newFollow;
pANTLR3_PARSER parser;
pANTLR3_TREE_PARSER tparser;
pANTLR3_INT_STREAM is;
switch (recognizer->type)
{
case ANTLR3_TYPE_PARSER:
parser = (pANTLR3_PARSER) (recognizer->super);
tparser = NULL;
is = parser->tstream->istream;
break;
case ANTLR3_TYPE_TREE_PARSER:
tparser = (pANTLR3_TREE_PARSER) (recognizer->super);
parser = NULL;
is = tparser->ctnstream->tnstream->istream;
break;
default:
fprintf(stderr, "Base recognizerfunction recover called by unknown paresr type - provide override for this function\n");
return ANTLR3_FALSE;
break;
}
newFollow = NULL;
if (follow == NULL)
{
/* The follow set is NULL, which means we don't know what can come
* next, so we "hit and hope" by just signifying that we cannot
* recover, which will just cause the next token to be consumed,
* which might dig us out.
*/
return ANTLR3_FALSE;
}
/* We have a bitmap for the follow set, hence we can compute
* what can follow this grammar element reference.
*/
if (follow->isMember(follow, ANTLR3_EOR_TOKEN_TYPE) == ANTLR3_TRUE)
{
/* First we need to know which of the available tokens are viable
* to follow this reference.
*/
viableToksFollowingRule = recognizer->computeCSRuleFollow(recognizer);
/* Knowing that, we can or in the follow set
*/
newFollow = follow->or(follow, viableToksFollowingRule);
/* Remove the EOR token, which we do not wish to compute with
*/
newFollow->remove(follow, ANTLR3_EOR_TOKEN_TYPE);
viableToksFollowingRule->free(viableToksFollowingRule);
/* We now have the computed set of what can follow the current token
*/
follow = newFollow;
}
/* We can now see if the current token works with the set of tokens
* that could follow the current grammar reference. If it looks like it
* is consistent, then we can "insert" that token by not throwing
* an exception and assumimng that we saw it.
*/
if ( follow->isMember(follow, is->_LA(is, 1)) == ANTLR3_TRUE)
{
/* report the error, but don't cause any rules to abort and stuff
*/
recognizer->reportError(recognizer);
if (newFollow != NULL)
{
newFollow->free(newFollow);
}
recognizer->error = ANTLR3_FALSE;
recognizer->failed = ANTLR3_FALSE;
return ANTLR3_TRUE; /* Success in recovery */
}
if (newFollow != NULL)
{
newFollow->free(newFollow);
}
/* We could not find anything viable to do, so this is going to
* cause an exception.
*/
return ANTLR3_FALSE;
}
/** Attempt to recover from a single missing or extra token.
*
* EXTRA TOKEN
*
* LA(1) is not what we are looking for. If LA(2) has the right token,
* however, then assume LA(1) is some extra spurious token. Delete it
* and LA(2) as if we were doing a normal match(), which advances the
* input.
*
* MISSING TOKEN
*
* If current token is consistent with what could come after
* ttype then it is ok to "insert" the missing token, else throw
* exception For example, Input "i=(3;" is clearly missing the
* ')'. When the parser returns from the nested call to expr, it
* will have call chain:
*
* stat -> expr -> atom
*
* and it will be trying to match the ')' at this point in the
* derivation:
*
* => ID '=' '(' INT ')' ('+' atom)* ';'
* ^
* match() will see that ';' doesn't match ')' and report a
* mismatched token error. To recover, it sees that LA(1)==';'
* is in the set of tokens that can follow the ')' token
* reference in rule atom. It can assume that you forgot the ')'.
*
* May need ot come back and look at the exception stuff here, I am assuming
* that the exception that was passed in in the java implementation is
* sotred in the recognizer exception stack. To 'throw' it we set the
* error flag and rules can cascade back when this is set.
*/
static void
recoverFromMismatchedToken (pANTLR3_BASE_RECOGNIZER recognizer, ANTLR3_UINT32 ttype, pANTLR3_BITSET follow)
{
pANTLR3_PARSER parser;
pANTLR3_TREE_PARSER tparser;
pANTLR3_INT_STREAM is;
switch (recognizer->type)
{
case ANTLR3_TYPE_PARSER:
parser = (pANTLR3_PARSER) (recognizer->super);
tparser = NULL;
is = parser->tstream->istream;
break;
case ANTLR3_TYPE_TREE_PARSER:
tparser = (pANTLR3_TREE_PARSER) (recognizer->super);
parser = NULL;
is = tparser->ctnstream->tnstream->istream;
break;
default:
fprintf(stderr, "Base recognizerfunction recoverFromMismatchedToken called by unknown paresr type - provide override for this function\n");
return;
break;
}
/* If the next token after the one we are looking at in the input stream
* is what we are looking for then we remove the one we have discovered
* from the stream by consuming it, then consume this next one along too as
* if nothing had happened.
*/
if ( is->_LA(is, 2) == ttype)
{
/* Print out the error
*/
recognizer->reportError(recognizer);
/* Call resync hook (for debuggeres and so on)
*/
recognizer->beginResync(recognizer);
/* "delete" the extra token
*/
is->consume(is);
/* End resync hook
*/
recognizer->endResync(recognizer);
/* consume the token that the rule actually expected to get
*/
is->consume(is);
recognizer->error = ANTLR3_FALSE; /* Exception is not outstanding any more */
}
/* The next token (after the one that is current, is not the one
* that we were expecting, so the input is in more of an error state
* than we hoped.
* If we are able to recover from the error using the follow set, then
* we are hunky dory again and can move on, if we cannot, then we resort
* to throwing the exception.
*/
if (recognizer->recoverFromMismatchedElement(recognizer, follow) == ANTLR3_FALSE)
{
recognizer->error = ANTLR3_TRUE;
recognizer->failed = ANTLR3_TRUE;
return;
}
}