About clearInsertRow() (connectivity and dbaccess module)

[Lionel Elie Mamane]

On Sat, Mar 30, 2013 at 08:07:15AM -0700, julien2412 [via Document Foundation Mail Archive] wrote:

> I noticed that a variable wasn't incremented in this loop:

>    1772 void OResultSet::clearInsertRow()
>    1773 {
>   ...
>    1778     for(sal_Int32 nPos = 0;aIter != aEnd;++aIter,++nPos)
>    1779     {
>    1780         ORowSetValueDecoratorRef& rValue = (*aIter);
>    1781         if ( rValue->isBound() )
>    1782         {
>    1783             (m_aRow->get())[nPos]->setValue( (*aIter)->getValue() );
>    1784         }
>    1785         rValue->setBound(nPos == 0);
>    1786         rValue->setModified(sal_False);
>    1787         rValue->setNull();
>    1788     }
>    1789 }

> See
> http://opengrok.libreoffice.org/xref/core/connectivity/source/drivers/file/FResultSet.cxx#1772

Which variable is not incremented? I don't see what is not
incremented; both aIter and nPos are incremented, by the statement
"++aIter,++nPos"

> I searched about "clearInsertRow" and found this method:
>    1638 void ORowSetCache::clearInsertRow()
>    1639 {
>    1640     // we don't unbound the bookmark column
>    1641     if ( m_aInsertRow != m_pInsertMatrix->end() && m_aInsertRow->is() )
>    1642     {
>    1643         ORowSetValueVector::Vector::iterator aIter = (*m_aInsertRow)->get().begin()+1;
>    1644         ORowSetValueVector::Vector::iterator aEnd = (*m_aInsertRow)->get().end();
>    1645         for(;aIter != aEnd;++aIter)
>    1646         {
>    1647             aIter->setBound(sal_False);
>    1648             aIter->setModified(sal_False);
>    1649             aIter->setNull();
>    1650         }
>    1651     }
>    1652 }
> http://opengrok.libreoffice.org/xref/core/dbaccess/source/core/api/RowSetCache.cxx#1638

> I wonder if it's normal that these functions are different. For example,
> should we consider a bookmark column in first method?

Well, one is in a specific driver and the other in a generic wrapper
/ cache, so the first may well make things specific to the underlying
database...

Specifically, well, the first "treats" the bookmark column (which is
always and by definition the column 0; the normal columns are numbered
starting from 1), and sets it to bound.

The second just skips the bookmark column.

I don't immediately see anything wrong with the difference, but I
cannot explain it either, so basically I'm not sure there.

> Then if it's ok, should we increment nPos in first one (...)

nPos is incremented.

-- 
Lionel




--
View this message in context: http://nabble.documentfoundation.org/About-clearInsertRow-connectivity-and-dbaccess-module-tp4047094p4047882.html
Sent from the Dev mailing list archive at Nabble.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/libreoffice/attachments/20130404/de9b1360/attachment.html>