We do a "SELECT * FROM table" just to fetch the primary key columns;
so reuse the same XResultSet to fetch all columns.
Else, we immediately issue a "SELECT * FROM table WHERE
primary_key=current_value" to read the other columns, which is
wasteful and particularly silly.
Commit 1ae17f5b03 already tried
to do that, but was essentially reverted piecewise because
it caused fdo#47520, fdo#48345, fdo#50372.
Commit c08067d6da94743d53217cbc26cffae00a22dc3a thought it did that,
but actually reverted commit 1ae17f5b03.
This implementation fetches the whole current row and caches it in memory;
only one row is cached: when the current row changes, the cache contains
the new current row.
This could be problematic (wrt to memory consumption) if the current
row is big (e.g. with BLOBs) and nobody is interested in the data
anyway (as would often be the case with BLOBs). Note that because of
our "SELECT *", the driver most probably has it in memory already
anyway, so we don't make the situation that much worse.
This could be incrementally improved with a heuristic of not
preemptively caching binary data (and also not LONGVARCHAR / TEXT /
MEMO / ...); a getFOO on these columns would issue a specific "SELECT
column FROM table WHERE primary_key=current_value" each time.
The *real* complete fix to all these issues would be to not do "SELECT
*" at all. Use "SELECT pkey_col1, pkey_col2, ..." when we are only
interested in the key columns. As to data, somehow figure out which
columns were ar interested in and "SELECT" only these (and maybe only
those with "small datatype"?). Interesting columns could be determined
by our caller (creator) as an argument to our constructor, or some
heuristic (no binary data, no "big" unbound data).
Also be extra smart and use *(m_aKeyIter) when getFOO is called
on a column included in it (and don't include it in any subsequent
SELECT).
However, there are several pitfalls.
One is buggy drivers that give use column names of columns that we
cannot fetch :-| Using "SELECT *" works around that because the driver
there *obviously* gives us only fetchable columns in the result.
Another one is the very restrictive nature of some database access
technologies. Take for example ODBC:
- Data can be fetched only *once* (with the SQLGetData interface;
bound columns offer a way around that, but that's viable only for
constant-length data, not variable-length data).
This could be addressed by an intelligent & lazy cache.
- Data must be fetched in increasing order of column number
(again, this is about SQLGetData).
This is a harder issue. The current solution has the nice advantage
of completely isolating the rest of LibO from these restrictions.
I don't currently see how to cleanly avoid (potentially
unnecessarily) caching column 4 if we are asked for column 3 then
column 5, just in case we are asked for column 4 later on, unless
we issue a specific "SELECT column4" later. But the latter would be
quite expensive in terms of app-to-database roudtripe times :-( and
thus creates another performance issue.
Change-Id: I999b3f8f0b8a215acb390ffefc839235346e8353
* commit 'ooo/DEV300_m101': (76 commits)
masterfix DEV300: #i10000# usage of L10N build_type
removetooltypes01: Fix build problems after rebase to DEV300m99 in basctl, cui, reportdesign, sw
hr75: #i116747#: remove obsolete copyright notices
gnumake3: remove comphelper version; fix including extract.hxx
removetooltypes01: #i112600# Replace missing tools types replacement
removetooltypes01: #i112600# remove tooltypes from reportdesign
removetooltypes01: #i112600# remove tooltypes from dbaccess
undoapi: don't use deprecated JUnit API, this seems to fail now (but didn't before the rebase)
undoapi: those two classes do not belong into JAVATESTFILES, but JAVAFILES only
undoapi: (I|Sfx)UndoManager now working with size_t instead of USHORT
gridsort: give the UnoControl(Model/Base) classes a ctor taking a service factory, so we have access to the factory which created us, and don't need to resort to the process'es service factory
dba34b: #109956# notify column value in correct order
dba34b: #i114309# set default values for bit boolean and reset when we are on the insertrow
undoapi: assert unknown slots in ImplInvalidateFeature
dba34b: #i110907# check if design handle event
dba34b: #i109956# notify column values when row is refreshed
dba34b: #i115753# add test again
undoapi: renamed Undo.* to RptUndo.*, to not confuse a certain IDE with different same-named files in the source tree ...
undoapi: migrated report designer to use an own, model-bound UndoManager, instead of the UndoManager formerly bound to the controller
undoapi: split OSingleDocumentController into DBSubComponentController (which the report design's controller will be derived from) and the still-so-named OSingleDocumentController. Module reportdesign does not compile with this change, this is yet to come.
...
Conflicts:
dbaccess/inc/IEnvironment.hxx
dbaccess/inc/dataview.hxx
dbaccess/qa/complex/dbaccess/RowSet.java
dbaccess/source/core/api/CacheSet.cxx
dbaccess/source/core/api/KeySet.hxx
dbaccess/source/core/api/OptimisticSet.hxx
dbaccess/source/core/api/RowSet.cxx
dbaccess/source/core/api/RowSetCache.cxx
dbaccess/source/ext/adabas/Acomponentmodule.cxx
dbaccess/source/ext/adabas/Aservices.cxx
dbaccess/source/filter/migration/cfgimport.cxx
dbaccess/source/filter/migration/cfgimport.hxx
dbaccess/source/filter/migration/cfgservices.cxx
dbaccess/source/filter/migration/makefile.mk
dbaccess/source/inc/cfg_reghelper.hxx
dbaccess/source/sdbtools/misc/sdbt_services.cxx
dbaccess/source/shared/cfg_reghelper.cxx
dbaccess/source/shared/cfgstrings.cxx
dbaccess/source/shared/registrationhelper.cxx
dbaccess/source/ui/app/AppController.cxx
dbaccess/source/ui/app/AppDetailPageHelper.cxx
dbaccess/source/ui/app/AppDetailPageHelper.hxx
dbaccess/source/ui/app/AppDetailView.cxx
dbaccess/source/ui/app/AppIconControl.cxx
dbaccess/source/ui/app/AppSwapWindow.cxx
dbaccess/source/ui/browser/brwctrlr.cxx
dbaccess/source/ui/browser/dataview.cxx
dbaccess/source/ui/browser/unodatbr.cxx
dbaccess/source/ui/control/dbtreelistbox.cxx
dbaccess/source/ui/control/opendoccontrols.cxx
dbaccess/source/ui/control/tabletree.cxx
dbaccess/source/ui/dlg/ConnectionHelper.cxx
dbaccess/source/ui/dlg/ConnectionPage.src
dbaccess/source/ui/dlg/ExtensionNotPresent.cxx
dbaccess/source/ui/dlg/ExtensionNotPresent.hrc
dbaccess/source/ui/dlg/ExtensionNotPresent.src
dbaccess/source/ui/dlg/TablesSingleDlg.cxx
dbaccess/source/ui/dlg/adodatalinks.hxx
dbaccess/source/ui/dlg/dbadmin2.src
dbaccess/source/ui/dlg/dbadminsetup.src
dbaccess/source/ui/dlg/dlgattr.cxx
dbaccess/source/ui/dlg/dlgsize.src
dbaccess/source/ui/dlg/makefile.mk
dbaccess/source/ui/dlg/sqlmessage.cxx
dbaccess/source/ui/inc/ExtensionNotPresent.hxx
dbaccess/source/ui/inc/TokenWriter.hxx
dbaccess/source/ui/inc/imageprovider.hxx
dbaccess/source/ui/misc/DExport.cxx
dbaccess/source/ui/misc/HtmlReader.cxx
dbaccess/source/ui/misc/RowSetDrop.cxx
dbaccess/source/ui/misc/RtfReader.cxx
dbaccess/source/ui/misc/TokenWriter.cxx
dbaccess/source/ui/misc/imageprovider.cxx
dbaccess/source/ui/misc/linkeddocuments.cxx
dbaccess/source/ui/misc/singledoccontroller.cxx
dbaccess/source/ui/querydesign/ConnectionData.hxx
dbaccess/source/ui/querydesign/JoinTableView.cxx
dbaccess/source/ui/querydesign/SelectionBrowseBox.cxx
dbaccess/source/ui/querydesign/TableConnectionData.cxx
dbaccess/source/ui/querydesign/TableWindowListBox.cxx
dbaccess/source/ui/relationdesign/RelationController.cxx
dbaccess/source/ui/relationdesign/RelationTableView.cxx
dbaccess/source/ui/tabledesign/table.src
reportdesign/inc/RptPage.hxx
reportdesign/qa/complex/reportdesign/ReportDesignerTest.java
reportdesign/source/core/api/ReportDefinition.cxx
reportdesign/source/core/sdr/ReportUndoFactory.cxx
reportdesign/source/core/sdr/UndoEnv.cxx
reportdesign/source/filter/xml/xmlservices.cxx
reportdesign/source/ui/dlg/GroupsSorting.cxx
reportdesign/source/ui/dlg/Navigator.cxx
reportdesign/source/ui/inspection/metadata.cxx
reportdesign/source/ui/misc/RptUndo.cxx
reportdesign/source/ui/misc/UITools.cxx
reportdesign/source/ui/misc/rptuiservices.cxx
reportdesign/source/ui/report/DesignView.cxx
reportdesign/source/ui/report/FixedTextColor.cxx
reportdesign/source/ui/report/ReportController.cxx
reportdesign/source/ui/report/ReportSection.cxx
reportdesign/source/ui/report/SectionView.cxx
reportdesign/source/ui/report/SectionWindow.cxx
reportdesign/source/ui/report/StartMarker.cxx
reportdesign/source/ui/report/ViewsWindow.cxx
reportdesign/source/ui/report/dlgedfunc.cxx
reportdesign/source/ui/report/report.src
2009-06-29 20:53:25 +0200 fs r273484 : #i103138# Rectangle conversion
2009-06-29 20:51:50 +0200 fs r273483 : #i103138# yet more refactoring, now also setting the proper zoom level at the proper point in time
2009-06-29 13:40:26 +0200 fs r273470 : added svn:ignore to ignore output paths
2009-06-29 10:08:54 +0200 fs r273455 : #i103138#
refactored the code for positioning/zooming the control
Basically, we now allow adjustControlGeometry_throw (formerly known as positionControl_throw and setControlZoom) to
take an additional ViewTransformation parameter, describing the transformation to obtain the actual
control position/size. Consequently, positionControl itself also allows for a ViewTransformation parameter.
This has become necessary since during painting, the device which we created our control for might not necessarily
have a proper MapMode set. In this case, if we would use this map mode for calculating the control's position/size,
this would lead to wrong results.
Note that this problem was introduced by the fix for #i101398#: During the fix, we postponed the control creation
to a later time (when it is really needed). At this later time, the MapMode at the device is broken, at the earlier
time where we formerly crearted the control (createPrimitive2DSequence), it is not yet broken.
Whether or not the MapMode is defined as "broken" might depend on one's point of view, however ...
I consider it broken, since:
- we need the map mode to obtain the proper zoom level, which is to be forwarded to the control
- there are scenarios where the MapMode is *not* set to MAP_PIXEL (in those scenarios, everything works
fine), and there are scenarios where it *is* set to MAP_PIXEL (in those the bug 103138 appears).
It somehow feels wrong that one cannot rely on the device's map mode this way, but on the other hand
one has no possibility to obtain the current zoom by other means.
Note that one issue (still to be submitted) is left: In the page pane of a Draw/Impress document, controls
have a wrong text size. This is because in this pane, the above-mentioned "broken" map mode is used,
which means the controls have a zoom of "1:1" set, which is wrong here.
2009-06-25 13:41:35 +0200 msc r273380 : #100000# the tabs changed die to new properties
2009-06-24 12:42:40 +0200 msc r273330 : #102082# remove issue warning
2009-06-22 10:43:14 +0200 fs r273201 : createPrimitive2DSequence: care for being disposed
2009-06-18 12:35:13 +0200 oj r273109 : #i102305# make nooptfiles for gcc
2009-06-17 12:14:37 +0200 oj r273056 : #i102305# fix for linux
2009-06-17 07:20:22 +0200 oj r273046 : #i102305# move ValueTransfer into the for loop to avoid a crash under Linux
2009-06-17 07:17:28 +0200 oj r273045 : #i102305# use varchar
2009-06-15 14:11:27 +0200 fs r272983 : added since tag
2009-06-15 12:11:39 +0200 oj r272973 : #i102305# SAL_DLLPUBLIC_EXPORT inserted
2009-06-15 11:08:53 +0200 fs r272969 : #i10000#
2009-06-15 09:25:13 +0200 fs r272963 : merging fix for P1 issue #i102701#
2009-06-11 11:31:24 +0200 fs r272858 : #i10000# copied the fix which before the rebase was done in ../dialog/macropg.src
2009-06-11 09:38:14 +0200 fs r272846 : CWS-TOOLING: rebase CWS dba32c to trunk@272827 (milestone: DEV300:m50)
2009-06-02 09:53:10 +0200 fs r272483 : #i10000#
2009-05-29 15:55:03 +0200 fs r272465 : #i100818#
2009-05-29 12:58:43 +0200 fs r272452 : don't apply comphelper::getString on possibly VOID any
2009-05-29 10:38:35 +0200 oj r272437 : #i101519# handle where condition
2009-05-29 09:53:39 +0200 fs r272434 : #i100818# call into releaseStubs /without/ locked GlobalMutex
2009-05-28 07:53:44 +0200 oj r272375 : #i101369# parse tree changed
2009-05-27 14:53:36 +0200 fs r272347 : #i10000#
2009-05-27 09:29:15 +0200 oj r272327 : #i101626# check for double before hard cast
2009-05-27 09:13:58 +0200 oj r272326 : #i101626# handle void correctly
2009-05-27 08:04:39 +0200 oj r272321 : #i102256# wrong method signature used
2009-05-27 07:55:52 +0200 oj r272320 : #i101519# look up parameter typ if used in function
2009-05-27 06:49:07 +0200 oj r272319 : #i101519# set parameter from rowset as well
2009-05-26 13:30:56 +0200 oj r272297 : #i101987# impl XBatchExecution
2009-05-26 12:44:34 +0200 oj r272293 : #i101700# check if group is not set
2009-05-26 12:16:53 +0200 oj r272290 : #i101369# resolved some reduce7reduce problems with boolean_term and search_condition
2009-05-26 12:12:42 +0200 oj r272289 : #i101369# fix for or on one line criteria
2009-05-25 16:02:25 +0200 fs r272257 : #i999704# +PROPERTY_MOUSE_WHEEL_BEHAVIOR
2009-05-25 16:01:55 +0200 fs r272256 : merging the changes from CWS dba32b herein
2009-05-25 15:49:57 +0200 fs r272254 : #i999704#
2009-05-25 15:32:57 +0200 fs r272252 : #i99704# grid columns also to respect the MouseWheelBehavior property
2009-05-25 15:23:43 +0200 fs r272251 : don't pass empty Anys to ::comphelper::getString
2009-05-25 14:48:43 +0200 fs r272248 : merged changes from CWS dba32b herein
2009-05-25 14:44:40 +0200 fs r272247 : #i99704# support new MouseWheelBehavior property
2009-05-25 14:43:18 +0200 fs r272246 : #i99704# WheelWithoutFocus (peer property) superseded by MouseWheelBehavior (model property)
2009-05-25 14:41:03 +0200 fs r272245 : #i99704# no need to set the mouse wheel behavior at the peer, this is now a model property, having the right default
2009-05-25 14:39:31 +0200 fs r272243 : removed dead import
2009-05-25 14:35:36 +0200 fs r272242 : the new EnableVisible doesn't make sense for grid columns
2009-05-25 14:34:33 +0200 fs r272241 : #i99704# +MouseWheelBehavior - allow to enable/disable the mouse wheel for the control, or make it focus-dependent
2009-05-25 14:26:11 +0200 fs r272240 : #i99704# change MouseSettings wheel flag (NoWheelActionWithoutFocus) to a three-state option, allowing to completely ignore the mouse wheel
2009-05-23 21:35:59 +0200 fs r272213 : localize 'sub component opened/closed' event
2009-05-22 21:42:47 +0200 fs r272211 : #i102003#
2009-05-22 21:42:20 +0200 fs r272210 : grammar
2009-05-22 21:36:10 +0200 fs r272209 : #i102140# load only once, not twice, and show error messages during loading (and during any form action, that is) asynchronously
2009-05-22 21:35:11 +0200 fs r272208 : #i102140# +clear
2009-05-22 14:50:30 +0200 fs r272194 : #i102139# for newly created DB docs, set the MacroExecutionMode to USE_CONFIG
2009-05-22 12:03:42 +0200 fs r272180 : #i88878#
provided by noel.power@novell.com
implement a visibility property (EnableVisible) for toolkit controls, and usage in forms and UNO dialogs
2009-05-15 15:37:31 +0200 fs r271942 : #i100671# corrected some @since tags, so autodoc has better chances of correctly reading them
2009-05-15 15:33:11 +0200 fs r271940 : don't call comphelper::getFOO for VOID values
2009-05-15 15:08:31 +0200 fs r271937 : includes
2009-05-15 13:39:22 +0200 fs r271934 : #i101398# createPrimitive2DSequence: when we already have a control, use the old code. In particular, call positionControlForPaint
2009-05-15 12:33:48 +0200 fs r271933 : make the geometry a part of the ControlPrimitive2D's identity
2009-05-15 10:15:44 +0200 fs r271928 : #i10000#
2009-05-14 20:55:38 +0200 fs r271921 : #i101398# don't reuse the name PRIMITIVE_ID_CONTROLPRIMITIVE2D, make the name of our own ControlPrimitive2D unique
2009-05-14 20:55:31 +0200 fs r271920 : #i101398# don't reuse the name PRIMITIVE_ID_CONTROLPRIMITIVE2D, make the name of our own ControlPrimitive2D unique
2009-05-14 20:23:23 +0200 fs r271919 : #i101622#
2009-05-14 16:04:38 +0200 fs r271898 : don't use comphelper::getInt32 on voids
2009-05-14 16:04:12 +0200 fs r271897 : merge fix for issue whose number just slipped my memory ... (originally fixed in CWS dba32b)
2009-05-14 15:36:55 +0200 fs r271895 : merging changes from DEV300:m48
2009-05-07 14:43:19 +0200 fs r271670 : #i101477#
2009-05-07 14:37:30 +0200 fs r271668 : #i101477#
2009-05-07 09:27:30 +0200 oj r271628 : #i101343# remove pch
2009-05-06 09:36:02 +0200 fs r271568 : getFoo: diagnostics
2009-05-04 09:23:06 +0200 oj r271438 : CWS-TOOLING: rebase CWS dba32c to trunk@271427 (milestone: DEV300:m47)
2009-04-29 23:18:13 +0200 fs r271394 : #i101398# use a dedicated 2DPrimitive for UNO Controls, which is able to provide the B2DRange *without* actually creating the control
2009-04-29 13:52:25 +0200 fs r271366 : #i101308#
2006/06/27 11:54:04 fs 1.24.2.2: RESYNC: (1.24-1.25); FILE MERGED
2006/06/16 14:16:04 fs 1.24.2.1: during #i51143#: replaced TTypeDefaultValuePair/TPositionTypePair with struct with named members - this is much more readable than ::std::pair< ..., ::std::pair< ..., ... > >
2004/11/09 14:29:15 oj 1.18.96.2: #i30004# use new singlequerycomposer
2004/10/18 10:40:30 oj 1.18.96.1: #i30004# don't use params from the rowset any longer
