ORCL PL/SQL grammar converted to case-insensitive; lexer simplified;

[OleksiiKovalov]

What does this PR do?

1. Makes grammar case insensitive
2. simplifies lexer by removes literals which are not directly used by production rules (except lists of "literals which can be used as identifiers")
   Why?
3. according to https://docs.oracle.com/cd/A97630_01/appdev.920/a96624/02_funds.htm#10519 "PL/SQL is not case sensitive, so lower-case letters are equivalent to corresponding upper-case letters except within string and character literals." so it is good to follow this rule. also, using "uppercase charstream" will break case of strings/literals - it may be vital for code analysis
4. code size is less and slightly faster. much faster compilation.
   not-used literals removed;
   lexer size 1056179 -> 343687 bytes
   parser size 3096431 -> 2752576 bytes

How is it checked?
using standard post-commit checks/ci

[OleksiiKovalov]

hi @KvanTTT , @kaby76
i need to re-generate .tree file for one of the example script ( sql/plsql/examples/create_procedure.sql.tree )
How can I do that? Please advise
thanks!

[kaby76]

hi @KvanTTT , @kaby76 i need to re-generate .tree file for one of the example script ( sql/plsql/examples/create_procedure.sql.tree ) How can I do that? Please advise thanks!

Yes, that's a good question. The way I normally do this is to generate a parser using trgen -t Java and call the executable with the -tree option. But, I don't know the official way to generate the .tree files.

[OleksiiKovalov]

The way I normally do this is to generate a parser using trgen -t Java and call the executable with the -tree option

Works for me, thanks!

[OleksiiKovalov]

hi @KvanTTT ,
please take a look

[KvanTTT]

Also postponed until status resolving of case insensitivity feature.

[KvanTTT]

options { caseInsensitive=true; } will be available since the next version of ANTLR. That's why such transformation will not be needed.