Is this an "acceptable" parser change

[Mark Rotteveel]

Trying to fix https://github.com/FirebirdSQL/firebird/issues/8182, I
made the following change in the Firebird 5 parse.y:

From:
```
%type <exprNode> in_predicate_value
in_predicate_value
: table_subquery { $$ = $1; }
| '(' value_list ')' { $$ = $2; }
;

%type <selectExprNode> table_subquery
table_subquery
: '(' column_select ')' { $$ = $2; }
;
```
to:
```
%type <exprNode> in_predicate_value
in_predicate_value
: table_subquery { $$ = $1; }
| '(' value_list ')' { $$ = $2; }
;

%type <selectExprNode> table_subquery
table_subquery
: '(' column_select ')' { $$ = $2; }
| '(' table_subquery ')' { $$ = $2; }
;
```

This seems to work, but increases the shift/reduce conflicts from 69 to 70.

Is that an acceptable trade-off, or am I maybe missing an alternative
solution?
--
Mark Rotteveel

[Adriano dos Santos Fernandes]

Yes.


Adriano

[Mark Rotteveel]

Thanks, I'll commit to v5 and master tomorrow (and maybe v4 and v3, I'll
need to double check some things there).

Mark

--
Mark Rotteveel

[Dimitry Sibiryakov]

'Mark Rotteveel' via firebird-devel wrote 12.06.2025 17:06:
> I'll commit to v5 and master tomorrow (and maybe v4 and v3, I'll need to double
> check some things there).

What is a reason to commit this change into a stable branch?

--
WBR, SD.

[Mark Rotteveel]

Fixing the bug I linked in the initial message, which can bite people
using Hibernate.
--
Mark Rotteveel

[Dimitry Sibiryakov]

'Mark Rotteveel' via firebird-devel wrote 13.06.2025 12:48:
> Fixing the bug I linked in the initial message, which can bite people using
> Hibernate.

I thought that it is a missing feature, not a bug.

--
WBR, SD.