Hello,
[ Sorry for the second email, but I wanted to
separate this ]
Would a...
sqlite3_begin(sqlite3 *db, int type)
sqlite3_end( sqlite3 *db )
be helpful for people counting transactions
in a thread environment?
The only difference to doing the 'transaction
level' counting yourself is that those
functions would use the structure 'sqlite3 *db'
to keep track of if a tranaction is in progress
or not.
The begin and end functions could be configured
to begin/end the transaction *if* there isn't/is
one.
ie.
func1()
{
sqlite3_begin(...)
func2();
DoSQLStuff...
sqlite3_end(...)
}
func2()
{
sqlite3_begin(...)
DoSQLStuff...
sqlite3_end(...)
}
...would work in a threaded environment since
each thread as its own copy of 'sqlite3 *db'
anyway. func2() would also work as expected
whether called in or outside func1().
Dumb idea?
Do I get get to wear the dunce hat? :)
-
Kervin
