Королев Илья wrote:
> В сообщении от 17 октября 2006 16:04 Warren Young
> написал(a):
>> Королев Илья wrote:
>>> Why we can't just drop away static names[] ( and table)?
>> How would SSQLS build queries without a table of the field names?
>
> like this:
>
> template <class Manip> std::ostream& operator << (std::ostream&
> s, const
> DB_AdvBase_field_list<Manip>& obj) {
> s << obj.manip << "user_id" << obj.delem;
> s << obj.manip << "position" << obj.delem;
> s << obj.manip << "ts" << obj.delem;
> s << obj.manip << "pads";
> return s;
> }
>
> IMHO in such way we don't need names[] at all.
Well, as a static example, that's interesting, but I'm not going to
rebuild custom.pl to implement it. If I were to spend a lot of time on
reworking SSQLS it would be to completely replace it with a different
mechanism. The current macro-based mechanism is hard to understand and
manipulate, so when we do run into problems, they tend not to get addressed.
This topic's come up before. I proposed defining the structures in XML
or some other such language and create a preprocessor that translates
that into header files defining the structures. This would be similar
to the way things like IDL work in CORBA and COM. The idea fizzled
because the driving force behind it at that time was Borland C++
incompatibility, but that problem's taken care of itself, leaving my
solution unnecessary.
You're welcome to tackle it, if you like.
>>> The only advantage IMHO is that we can change names[](table) on the
>>> fly.
>> Every single table in my own use of MySQL++ has a different name than
>> the corresponding SSQLS. I am unwilling to give up that feature.
>
> Explain please.
I mean that I will define an SSQLS called something like
SQL_MyTableName, where the actual MySQL table name is something else,
like my_table_name. There are several reasons I do this, but it's not
important to argue them here. The important point is, I and others
depend on being able to have C++ structure names that are different from
the corresponding SQL table names, yet still bind them together via SSQLS.
> In such case IMHO it's possible to declare "TableName" in other way.
> #define TABLE_NAME "TableName"
> Q << "SELECT " << obj.field_list() << " FROM " <<
> TABLE_NAME;
Why would I want to do that? The whole point of SSQLS is to reduce the
amount of SQL statements in my C++ code. This is a regression.
