Author: wyoung
Date: Thu Nov 29 18:17:32 2007
New Revision: 1931
URL: http://svn.gna.org/viewcvs/mysqlpp?rev=1931&view=rev
Log:
Removed "basic" variants of SSQLS macros. They've been unofficially
deprecated for a long time now by virtue of not having any examples that
use them, and being marginalized in the userman.
Modified:
trunk/Wishlist
trunk/doc/userman/userman.dbx
trunk/lib/custom.pl
Modified: trunk/Wishlist
URL:
http://svn.gna.org/viewcvs/mysqlpp/trunk/Wishlist?rev=1931&r1=1930&r2=1931&view=diff
==============================================================================
--- trunk/Wishlist (original)
+++ trunk/Wishlist Thu Nov 29 18:17:32 2007
@@ -75,9 +75,6 @@
boilerplate, which leaf SSQLSes derive from. This should in turn
allow template functions like Query::insert<T> to become regular
member functions, taking a reference to the SSQLS base class.
-
- o Remove "basic" variants of SSQLS creation macros, and remove
- userman sections covering them.
o Either add quote_force and similar manipulators, or remove the
'r' and 'R' modifiers in template queries. As it stands, you
Modified: trunk/doc/userman/userman.dbx
URL:
http://svn.gna.org/viewcvs/mysqlpp/trunk/doc/userman/userman.dbx?rev=1931&r1=1930&r2=1931&view=diff
==============================================================================
--- trunk/doc/userman/userman.dbx (original)
+++ trunk/doc/userman/userman.dbx Thu Nov 29 18:17:32 2007
@@ -1863,32 +1863,28 @@
Structures</title>
<para>Up to this point, we haven't been using all of the
- features in the SSQLS structures we've been generating. We
- could have used the <symbol>sql_create_basic_*</symbol>
- macros instead, which would have worked just as well for
- what we've seen so far, and the generated code would have
- been smaller.</para>
-
- <para>Why is it worth ignoring the "basic" variants of these
- macros, then? Consider this:</para>
+ features in the SSQLS structures we've been generating.
+ What else can we do with SSQLSes? Consider this:</para>
<programlisting>
query.insert(s);</programlisting>
- <para>This does exactly what you think it does: it inserts
- 's' into the database. This is possible because a standard
- SSQLS has functions that the query object can call to get
- the list of fields and such, which it uses to build an
- insert query. <methodname>query::update()</methodname> and
- <methodname>query::replace()</methodname> also rely on this
- SSQLS feature. A basic SSQLS lacks these functions.</para>
-
- <para>Another feature of standard SSQLSes you might find a use
- for is changing the table name used in queries. By default,
- the table in the MySQL database is assumed to have the same
- name as the SSQLS structure type. But if this is inconvenient,
- you can globally change the table name used in queries like
- this:</para>
+ <para>This does exactly what you think it does: it builds
+ an <command>INSERT</command> query to insert the contents
+ of <varname>s</varname> into the database. You have only to
+ call <methodname>query::execute()</methodname> to actually
+ insert it. This is possible because SSQLSes have functions that
+ <classname>Query</classname> can call to get the list of fields
+ and such, which it uses to build the <command>INSERT</command>
+ query. <methodname>query::update()</methodname> and
+ <methodname>query::replace()</methodname> also rely on these
+ features.</para>
+
+ <para>Another feature you might find a use for is changing
+ the table name used in queries. By default, the table in the
+ MySQL database is assumed to have the same name as the SSQLS
+ structure type. But if this is inconvenient, you can globally
+ change the table name used in queries like this:</para>
<programlisting>
stock::table() = "MyStockData";</programlisting>
@@ -1898,21 +1894,18 @@
<sect2>
<title>Using an SSQLS in Multiple Modules</title>
- <para>Unless you use one of the emasculated forms of SSQLS
- described below, you'll run into a problem when trying to
- use it in more than one module. The more advanced forms of
- SSQLS include a few static data members to hold information
- common to all structures of that type. (The table name and
- the list of field names.) A typical way to run into this
- is to put the SSQLS definition in a header file and then
- <userinput>#include</userinput> that in all the modules that
- need to use SSQLSes of that type. This results in each module
- defining its own copy of the static data members, which causes
- a multiply-defined symbol error at link time.</para>
-
- <para>The way around this is to add this before all but one
- of the places where you pull in the header definining the
- SSQLS:</para>
+ <para>It's convenient to define an SSQLS in a header file
+ so you can use it in multiple modules. You run into a bit of
+ a problem, though, because each SSQLS includes a few static
+ data members to hold information common to all structures of
+ that type. (The table name and the list of field names.) When
+ you <command>#include</command> that header in more than
+ one module, you get a multiply-defined symbol error at link
+ time.</para>
+
+ <para>The way around this is to add this before <emphasis>all
+ but one</emphasis> of the places where you pull in the header
+ definining the SSQLS:</para>
<programlisting>
#define MYSQLPP_SSQLS_NO_STATICS</programlisting>
@@ -1928,6 +1921,15 @@
program that logically "owns" each SSQLS, and it can pull in
the header definining the SSQLS directly, without suppressing
the static data members.</para>
+
+ <para>Note that due to a compiler limitation, you can't
+ use this feature with Visual C++ 2003. As instructed
+ in <filename>README.vc</filename>, you have to disable
+ this feature in order to get the SSQLS header files to
+ compile. Having done that, the SSQLS feature works fine as
+ long as you can live with using each structure type in a
+ single module. Visual C++ 2005 and newer don't suffer from
+ this limitation.</para>
</sect2>
@@ -2098,11 +2100,6 @@
string, item, "item", 1,
int, num, "quantity", 2,
double, weight, "weight", 3)</programlisting>
-
- <para>All three of these macro types have "basic" variants
- that work the same way. Again, basic SSQLSes lack the features
- necessary for automatic insert, update and replace query
- creation.</para>
</sect2>
Modified: trunk/lib/custom.pl
URL:
http://svn.gna.org/viewcvs/mysqlpp/trunk/lib/custom.pl?rev=1931&r1=1930&r2=1931&view=diff
==============================================================================
--- trunk/lib/custom.pl (original)
+++ trunk/lib/custom.pl Thu Nov 29 18:17:32 2007
@@ -362,32 +362,6 @@
// ---------------------------------------------------
---
my $out = <<"---";
-#define sql_create_basic_c_order_$i(NAME, CMP, CONTR, $parm_order)
-
- struct NAME;
-
- template <mysqlpp::sql_dummy_type dummy> int sql_compare_##NAME (const NAME
&, const NAME &);
-
- struct NAME {
-$defs
- NAME () {}
- NAME (const mysqlpp::Row &row);
- sql_compare_define_##CMP(NAME, $parmC)
- };
-
- template <mysqlpp::sql_dummy_type dummy>
- void populate_##NAME (NAME *s, const mysqlpp::Row &row) {
-$popul
- }
-
- inline NAME::NAME (const mysqlpp::Row &row)
- {populate_##NAME<mysqlpp::sql_dummy>(this, row);}
-
- sql_COMPARE__##CMP(NAME, $parmc )
----
- print OUT &prepare($out);
-
- $out = <<"---";
#define sql_create_complete_$i(NAME, CMP, CONTR, $parm_complete)
struct NAME;
@@ -891,9 +865,6 @@
#
print OUT << "---";
-#define sql_create_basic_$i(NAME, CMP, CONTR, $parm_simple_b) \\
- sql_create_basic_c_order_$i(NAME, CMP, CONTR, $parm_simple2c_b)
-
#define sql_create_$i(NAME, CMP, CONTR, $parm_simple) \\
sql_create_complete_$i(NAME, CMP, CONTR, $parm_simple2c) \\
_______________________________________________
Mysqlpp-commits mailing list
[email protected]
https://mail.gna.org/listinfo/mysqlpp-commits
