Bump the yacc pointer. #1804
Bump the yacc pointer. #1804
Conversation
petermattis
force-pushed
the
pmattis/sql-yacc
branch
from
e61b02c to
04b51ed
Compare
| @@ -129,7 +129,7 @@ check: | |||
| grep -vE '(declaration of err shadows|^vet: cannot process directory \.git)' | |||
| @echo "golint" | |||
| @! golint $(PKG) | \ | |||
| grep -vE '(\.pb\.go|embedded\.go|_string\.go|LastInsertId|sql\.y)' \ | |||
| grep -vE '(\.pb\.go|embedded\.go|_string\.go|LastInsertId|sql/parser/(yaccpar|sql\.y):)' \ | |||
There was a problem hiding this comment.
It is the "file" containing the generated parser code and is annotated in sql.go with:
//line yaccpar:1
|
LGTM |
Use the newer yacc to re-generate the sql grammar. Fixes #1801.
petermattis
force-pushed
the
pmattis/sql-yacc
branch
from
04b51ed to
8bfe748
Compare
|
LGTM. |
|
@tschottdorf The code that caused the panic has seen changes in the updated yacc code. I didn't look at the logs, but I'd guess that the similar problem go-fuzz found in the vitess parser resulted in changes to yacc. |
|
the posts in the sister thread in vitess sounded like "I dunno but it seems to work now". See vitessio/vitess#767. |
|
I was referring to the the git logs for |
Use the newer yacc to re-generate the sql grammar.
Fixes #1801.