[PATCH] Add Rust parser check for end of expression

[PATCH] Add Rust parser check for end of expression

[Tom Tromey]

I noticed that "print 5," passed in Rust -- the parser wasn't checking
that the entire input was used.  This patch fixes the problem.  This
in turn pointed out another bug in the parser, namely that it didn't
lex the next token after handling a string token.  This is also fixed
here.
---
 gdb/rust-parse.c                | 6 +++++-
 gdb/testsuite/gdb.rust/expr.exp | 2 ++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/gdb/rust-parse.c b/gdb/rust-parse.c
index 4006df7086b..7d7d882872c 100644
--- a/gdb/rust-parse.c
+++ b/gdb/rust-parse.c
@@ -271,7 +271,10 @@ struct rust_parser
   operation_up parse_entry_point ()
   {
     lex ();
-    return parse_expr ();
+    operation_up result = parse_expr ();
+    if (current_token != 0)
+      error (_("Syntax error near '%s'"), pstate->prev_lexptr);
+    return result;
   }
 
   operation_up parse_tuple ();
@@ -2020,6 +2023,7 @@ rust_parser::parse_atom (bool required)
 
     case STRING:
       result = parse_string ();
+      lex ();
       break;
 
     case BYTESTRING:
diff --git a/gdb/testsuite/gdb.rust/expr.exp b/gdb/testsuite/gdb.rust/expr.exp
index 0c445897338..bb0222bed4b 100644
--- a/gdb/testsuite/gdb.rust/expr.exp
+++ b/gdb/testsuite/gdb.rust/expr.exp
@@ -145,3 +145,5 @@ gdb_test "print 0x0 as fn(i64) -> ()" " = \\\(\\*mut fn \\\(i64\\\) -> \\\(\\\)\
 gdb_test "print r#" "No symbol 'r' in current context"
 
 gdb_test "printf \"%d %d\\n\", 23+1, 23-1" "24 22"
+
+gdb_test "print 5," "Syntax error near ','"
-- 
2.34.1

Re: [PATCH] Add Rust parser check for end of expression

[Pedro Alves]

On 2022-03-18 04:44, Tom Tromey wrote:
> I noticed that "print 5," passed in Rust -- the parser wasn't checking
> that the entire input was used.  This patch fixes the problem.  This
> in turn pointed out another bug in the parser, namely that it didn't
> lex the next token after handling a string token.  This is also fixed
> here.

OOC, is the lex-next-token bug something that is also covered by
the added test?  It doesn't seem to use strings.

Pedro Alves

> ---
>  gdb/rust-parse.c                | 6 +++++-
>  gdb/testsuite/gdb.rust/expr.exp | 2 ++
>  2 files changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/gdb/rust-parse.c b/gdb/rust-parse.c
> index 4006df7086b..7d7d882872c 100644
> --- a/gdb/rust-parse.c
> +++ b/gdb/rust-parse.c
> @@ -271,7 +271,10 @@ struct rust_parser
>    operation_up parse_entry_point ()
>    {
>      lex ();
> -    return parse_expr ();
> +    operation_up result = parse_expr ();
> +    if (current_token != 0)
> +      error (_("Syntax error near '%s'"), pstate->prev_lexptr);
> +    return result;
>    }
>  
>    operation_up parse_tuple ();
> @@ -2020,6 +2023,7 @@ rust_parser::parse_atom (bool required)
>  
>      case STRING:
>        result = parse_string ();
> +      lex ();
>        break;
>  
>      case BYTESTRING:
> diff --git a/gdb/testsuite/gdb.rust/expr.exp b/gdb/testsuite/gdb.rust/expr.exp
> index 0c445897338..bb0222bed4b 100644
> --- a/gdb/testsuite/gdb.rust/expr.exp
> +++ b/gdb/testsuite/gdb.rust/expr.exp
> @@ -145,3 +145,5 @@ gdb_test "print 0x0 as fn(i64) -> ()" " = \\\(\\*mut fn \\\(i64\\\) -> \\\(\\\)\
>  gdb_test "print r#" "No symbol 'r' in current context"
>  
>  gdb_test "printf \"%d %d\\n\", 23+1, 23-1" "24 22"
> +
> +gdb_test "print 5," "Syntax error near ','"

Re: [PATCH] Add Rust parser check for end of expression

[Tom Tromey]

>> I noticed that "print 5," passed in Rust -- the parser wasn't checking
>> that the entire input was used.  This patch fixes the problem.  This
>> in turn pointed out another bug in the parser, namely that it didn't
>> lex the next token after handling a string token.  This is also fixed
>> here.

Pedro> OOC, is the lex-next-token bug something that is also covered by
Pedro> the added test?  It doesn't seem to use strings.

Yes, if you remove that hunk there will be other regressions.

Tom

Re: [PATCH] Add Rust parser check for end of expression

[Tom Tromey]

>>>>> "Tom" == Tom Tromey <tom@tromey.com> writes:

Tom> I noticed that "print 5," passed in Rust -- the parser wasn't checking
Tom> that the entire input was used.  This patch fixes the problem.  This
Tom> in turn pointed out another bug in the parser, namely that it didn't
Tom> lex the next token after handling a string token.  This is also fixed
Tom> here.

I meant to push this before the gdb 12 branch was made.  I'm going to
check it in and also backport it to the release branch.

Tom