Re: [PATCH] [PUSHED: 3-5] horrible performance regression

Hi Lionel,

sigh, I have somehow missed this mail.

Lionel Elie Mamane píše v St 18. 07. 2012 v 19:37 +0200:
> On Tue, Jul 17, 2012 at 12:10:05AM +0200, Lionel Elie Mamane wrote:
> > Please cherry-pick 0cda6605844ef68e45db7a7c05cc4d09ef2bc49a
> > (http://cgit.freedesktop.org/libreoffice/core/commit/?id=0cda6605844ef68e45db7a7c05cc4d09ef2bc49a
> >  and patch also attached)
> > to libreoffice-3-6 in time for rc2.
>
> > I'll make a backport to libreoffice-3-5 later this week
>
> So, the sorry situation is that backporting this without also
> backporting e581bef6dfc03d0bab9de1485c6f6cdcd034d581 is pointless.
>
> That's more than I'm comfortable with on a stable branch, but OTOH, if
> we don't fix this, we'd leave a huge problem for users of embedded
> HSQLDB. So I reluctantly ask for backport to libreoffice-3-5.
>
> Additionally, e581bef6dfc03d0bab9de1485c6f6cdcd034d581 moderately
> intertwined with 773668c6ab0963f56f98270b29d595f5df7c4bb2, so I
> decided that porting e581bef6dfc03d0bab9de1485c6f6cdcd034d581 without
> 773668c6ab0963f56f98270b29d595f5df7c4bb2 is actually more risky than
> without it.
>
>
> So the backport of these three commits is attached as patches. I
> squashed with fixups that happened shortly after these commits.

I do not pretend that I understand everything. I do not see any obvious
problem. I agree that it is better to have the same code in 3.5 and 3.6
in this case => pushed to 3-5 branch, see
http://cgit.freedesktop.org/libreoffice/core/commit/?h=libreoffice-3-5&id=7f75bc39fc1f49fe0b8c5c550eb801c0974a53a2
http://cgit.freedesktop.org/libreoffice/core/commit/?h=libreoffice-3-5&id=6085d9cc065a0612ea10e29274774ee0a6548d3a
http://cgit.freedesktop.org/libreoffice/core/commit/?h=libreoffice-3-5&id=186366967e7d57e1bd539a43c7383405adb8a575


Note that I removed the hunk:

--- cut ---
--- a/connectivity/source/drivers/jdbc/PreparedStatement.cxx
+++ b/connectivity/source/drivers/jdbc/PreparedStatement.cxx
@@ -150,6 +150,7 @@ void SAL_CALL
java_sql_PreparedStatement::setString( sal_Int32 parameterIndex, c
 
 ::com::sun::star::uno::Reference< ::com::sun::star::sdbc::XResultSet >
SAL_CALL java_sql_PreparedStatement::executeQuery(  )
throw(::com::sun::star::sdbc::SQLException, ::com::sun::star::uno::RuntimeException)
 {
+    std::cerr << ::rtl::OUStringToOString( this->m_sSqlStatement,
RTL_TEXTENCODING_UTF8 ).getStr() << std::endl;
     m_aLogger.log( LogLevel::FINE, STR_LOG_EXECUTING_PREPARED_QUERY );
     ::osl::MutexGuard aGuard( m_aMutex );
     checkDisposed(java_sql_Statement_BASE::rBHelper.bDisposed);
--- cut ---

IMHO, it was just a debugging output.


Hmm, I think that it is too risky to push it into 3-5-6 for 3.5.6-rc2.
So, it needs to wait for 3.5.7. Well, we will at least get some feedback
from 3.6.0 users in the meantime.

Best Regards,
Petr