On Tue, May 10, 2005 at 11:10:31AM -0600, Warren Young wrote:
From my research, there is a problem in custom.pl, on line 603 (in 1.7.34)
.... there is a template member function that has no definition in the
rest of the header. Specifically:

template <class Manip>
NAME##_cus_equal_list<Manip> equal_list(cchar *d, cchar *c, Manip m,
std::vector<bool> *i) const;

Mark found some more.


This example tickles the bug, since equal_list(cchar*, vector<bool>*)
calls the above missing member function.

Below is Mark's patch, diff-ified, plus some warning fixes for out-of-order
constructor arguments (gcc complains about this regularly). I ran the simple
example which I also included in this patch. If Mark could retest this,
that would be great.

- Chris


Index: software/mysql++/examples/custom1.cpp
diff -u software/mysql++/examples/custom1.cpp:1.1.1.2 software/mysql++/examples/custom1.cpp:1.2
--- software/mysql++/examples/custom1.cpp:1.1.1.2 Thu Mar 3 23:27:00 2005
+++ software/mysql++/examples/custom1.cpp Tue May 10 16:08:02 2005
@@ -43,6 +43,17 @@
// this is storing the results into a vector of the custom struct
// "stock" which was created my the macro above.

+ {
+ Query query = con.query();
+ vector<bool> a(5, false);
+ a[stock_weight] = true;
+ a[stock_price] = true;
+ query << "select * from stock where " <<
+ res[0].equal_list(" and ", &a);
+ cout << "Test query:" << endl;
+ cout << query.preview() << endl;
+ }
+
cout.setf(ios::left);
cout << setw(17) << "Item"
<< setw(4) << "Num"
Index: software/mysql++/lib/custom.pl
diff -u software/mysql++/lib/custom.pl:1.7 software/mysql++/lib/custom.pl:1.9
--- software/mysql++/lib/custom.pl:1.7 Mon May 2 06:25:53 2005
+++ software/mysql++/lib/custom.pl Tue May 10 16:08:02 2005
@@ -376,10 +376,10 @@
const NAME##_cus_value_list<Manip>&); */
public:
const NAME *obj;
- cchar *delem;
- Manip manip;
std::vector<bool> *include;
bool del_vector;
+ cchar *delem;
+ Manip manip;
public:
~NAME##_cus_value_list () {if (del_vector) delete include;}
NAME##_cus_value_list (const NAME *o, cchar *d, Manip m, $cusparms11);
@@ -394,10 +394,10 @@
const NAME##_cus_field_list<Manip>&); */
public:
const NAME *obj;
- cchar *delem;
- Manip manip;
std::vector<bool> *include;
bool del_vector;
+ cchar *delem;
+ Manip manip;
public:
~NAME##_cus_field_list () {if (del_vector) delete include;}
NAME##_cus_field_list (const NAME *o, cchar *d, Manip m, $cusparms11);
@@ -772,6 +772,24 @@
}

template <class Manip>
+ inline NAME##_cus_value_list<Manip> NAME::value_list(cchar *d, Manip m,
+ std::vector<bool> *i) const {
+ return NAME##_cus_value_list<Manip> (this, d, m, i);
+ }
+
+ template <class Manip>
+ inline NAME##_cus_field_list<Manip> NAME::field_list(cchar *d, Manip m,
+ std::vector<bool> *i) const {
+ return NAME##_cus_field_list<Manip> (this, d, m, i);
+ }
+
+ template <class Manip>
+ inline NAME##_cus_equal_list<Manip> NAME::equal_list(cchar *d, cchar *c, Manip m,
+ std::vector<bool> *i) const {
+ return NAME##_cus_equal_list<Manip> (this, d, c, m, i);
+ }
+
+ template <class Manip>
inline NAME##_cus_value_list<Manip>
NAME::value_list(cchar *d, Manip m, sql_cmp_type /*sc*/) const {
sql_compare_type_def_##CMP (NAME, value, NUM);


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:

Thank you for the explanation.

Thanks again.

This is good to fix, but you should to do it by reordering the
initializer list, not by reordering the member declarations. I don't
think it's an ABI breakage to make the change as you have, but it is
needless.

I'd rather that you created a new example, rather than modify
custom1.cpp. Disk space is cheap; adding custom5.cpp won't be a problem.

I retract my wish for bare constants here. I didn't realize that
custom.h defined these constants for you. It will be good to document
this in an example.

Last but not least, can we get a DocBook comment for the new functions?
I'm going to request this increasingly, so you all might as well get
used to it. I'm not going to have all that hard work in the new
docs go to waste by letting them slip back into irrelevancy.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Tue, May 10, 2005 at 02:35:42PM -0600, Warren Young wrote:
I thought that at first too, but then I noticed there were other instances
of the same type of constructor, with variables in different order.
I figured if someone else was going to wade through custom.pl, it might help
to have constructor lists in the same order, for grep reasons.


Agreed.


Thanks to Mark for showing this in his example. I had no idea either.


I hope to do my share adding Doxygen comments wherever possible. Doxygen
isn't really possible in custom.pl, as far as I know.

For DocBook, I assume straight patches to doc/userman/userman.xml are
the preferred format? I haven't finished reading it all, so
don't feel qualified to change it yet. :-)

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

=20
Chris, Thanks for explaining that.

Chris, Thanks for doing this.

To all, sorry I have not presented things in the form that is expected.
I am new to this list and still learning how you would like to see
patches presented. I will catch on soon. Sorry.

I learned about the equal_list() function and the constants from section
7.7 of the old document. Personally I find them very useful, but I
believe they were removed in the new document. My suggestion would to
be to add back in the info and examples that were previously in section
7.7. Maybe just updated a bit or something.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:
Hmmm, okay, I buy that. I'm going to accept it provisionally, on the
condition that it doesn't break my existing binaries.

Instruction to self: open mouth, insert foot.

Sorry, I wasn't thinking when I wrote that. The sentiment is still
valid, though.

Yes, that's the source file. See the README file in that directory for
sources of information on DocBook. It really isn't all that different
from HTML, so if you know that, the learning curve should be pretty shallow.

I won't be asking as much for changes to the userman as to the refman.
"Voice" is important in the userman, so it there should be as few
fingers as practical in that pie. In the refman, it's just a recounting
of hard facts using a well-defined format, so there are few style
concerns with it.

Don't worry about documenting custom5. I'll handle that.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

On Wed, May 11, 2005 at 08:51:52AM -0400, Mark Merendino wrote:
No worries, you did the bulk of the brain work on that. :-)

You probably know this by now, but "diff -ruN mysql++-olddir mysql++-newdir"
is your friend.

I actually use CVS to keep track of all this, which uses less disk space.
Every time a new tarball comes out, I import it on the vendor branch like
this:

cvs -d /home/cvs import -m "Version 1.7.35" software/mysql++ \
mysql++ mysql++-1_7_35

Update HEAD like this:

cd mysql++
cvs update -PdA
cvs update -j HEAD -j mysql++-1_7_35 -Pd
cvs ci -m "merged 1.7.35 into HEAD"
cvs tag mysql++-1_7_35-merged

Then any changes I make can easily be tracked, and diffs can be produced
with rdiff.

vi lib/custom.pl
cvs ci -m "Mark Merendino's equal_list() fix"
cvs tag mysql++-1_7_35-equal_list
cvs rdiff -u -r mysql++-1_7_35 -r mysql++-1_7_35_equal_list \
$(cat CVS/Repository) > equal_list.patch

This may all change when Warren goes to a public subversion repository,
but it works well for me, and I have it all on my local machine.

- Chris


--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

Chris Frey wrote:

etc...

I've added this information to the HACKERS file. Thanks for taking the
time to write it up, Chris!

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw