C++ target fixes/changes/improvements

[ibre5041]

• Some Lexer fixes (old C-target code ported to C++)
• "Run-time headers" included (and modified)
• Bugfixes
• Parser return types named (shortened)
• ST warnings removed

[ibre5041]

I finally got familiar with git rebase command. The commit 4be2968 addresses you comment for 77d8914. This commit replaces three different commits, which change the same piece of the code.

[ibre5041]

AST support was removed form this pull request. Can pls review and approve this set of patches?
Also @ members stay in .cpp while @ context is in .hpp.

[ksgokul]

+1.. The changes look fine to me.

[ibre5041]

On 06/19/2013 07:27 PM, ksgokul wrote:

+1.. The changes look fine to me.

—
Reply to this email directly or view it on GitHub
#112 (comment).

Sam can you please merge the pull request 112?

thx Ivan

PS: Maybe I've found a reason why C++ target is sometimes slower that
Java one.
When using semantic predicates ANTLR generates code like this:
// PLSQLGuiLexer.g:1107:2: ( ( SQLPLUS_COMMAND_INTRODUCER
)=>sqlplus= SQLPLUS_COMMAND_INTRODUCER |)
{
int alt51=2;
{
int LA51_0 = this->LA(1);
if ( (this->msynpred(
antlr3::ClassForwarder<synpred1_PLSQLGuiLexer>() )) && (LA51_0 == 'A'))
{
alt51=1;
}
else if ( (this->msynpred(
antlr3::ClassForwarder<synpred1_PLSQLGuiLexer>() )) && (LA51_0 == 'B'))
{
alt51=1;
}
...
The order is wrong - msynpred is evaluated again and again. The same
applies to C target too. While Java target generates:
// PLSQLGuiLexer.g:1107:2: ( ( SQLPLUS_COMMAND_INTRODUCER
)=>sqlplus= SQLPLUS_COMMAND_INTRODUCER |)
int alt51=2;
int LA51_0 = input.LA(1);
if ( (LA51_0=='A') && (synpred1_PLSQLGuiLexer())) {
alt51=1;
}
else if ( (LA51_0=='B') && (synpred1_PLSQLGuiLexer())) {
alt51=1;
}
else if ( (LA51_0=='C') && (synpred1_PLSQLGuiLexer())) {
alt51=1;
}
...
So far I could not identify piece of code, which is responsible for this.

[ibre5041]

Can you please merge this pull request? It contains only bugfixes.

[sharwell]

We're working on releasing ANTLR 4.1 tomorrow. I'll look into this merge either tomorrow or Monday.

[ibre5041]

Ping

[parrt]

hahah! Almost got to it this weekend. dang it. thanks for ping.
T
On Sep 3, 2013, at 6:54 AM, ibre5041 wrote:

Ping

—
Reply to this email directly or view it on GitHub.