Romulo Goncalves wrote:
Update of /cvsroot/monetdb/sql/src/server In directory sc8-pr-cvs16:/tmp/cvs-serv14592/src/server
Modified Files: sql_schema.mx Log Message:
[since I spent (almost) all night and morning on compiling & testing, anyway:]
for the umpteenth time: make SQL compile with icc, again, by removind set but unused variable --- some more concentration during coding might save us quite some time ....
Romulo,
could you please check, whether the code still complies with you intention?
Index: sql_schema.mx =================================================================== RCS file: /cvsroot/monetdb/sql/src/server/sql_schema.mx,v retrieving revision 1.122 retrieving revision 1.123 diff -u -d -r1.122 -r1.123 --- sql_schema.mx 17 Apr 2007 12:27:07 -0000 1.122 +++ sql_schema.mx 18 Apr 2007 09:16:26 -0000 1.123 @@ -602,11 +602,10 @@ { char *tname = qname_table(qname); sql_schema *ss = cur_schema(sql); - sql_trigger * t= NULL;
if (!schema_privs(sql->role_id, ss)) return sql_error(sql, 02, "DROP TRIGGER: access denied for %s to schema ;'%s'", stack_get_string(sql, "current_user"), ss->base.name); - if ((t = mvc_bind_trigger(sql, ss, tname )) == NULL) + if (mvc_bind_trigger(sql, ss, tname ) == NULL) Why I should not use the t? I found similar examples in the code. Maybe
Stefan Manegold wrote: the the t was used after. I agree with this change I think it does affect the semantics of the function.
If you look at the code, the t was not used anywhere in that function. It was assigned a value, and that's it. That's what icc warns about (and rightly so!). Stefan's question boils down to the question, should that t have been used after it had been assigned, i.e. did you need to do anything with the result of the function call? If not, then the assignment and hence the variable was superfluous. -- Sjoerd Mullender