Issue 20826 - scrolling tooltip pain (patch)
Summary: scrolling tooltip pain (patch)
Status: CLOSED NOT_AN_OOO_ISSUE
Alias: None
Product: Writer
Classification: Application
Component: code (show other issues)
Version: OOo 1.1 RC5
Hardware: PC Linux, all
: P2 Trivial (vote)
Target Milestone: ---
Assignee: Oliver-Rainer Wittmann
QA Contact: issues@sw
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-10-07 10:38 UTC by mmeeks
Modified: 2003-12-22 05:37 UTC (History)
1 user (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments

Note You need to log in before you can comment on or make changes to this issue.
Description mmeeks 2003-10-07 10:38:19 UTC
Calling EndScrollHdl(pScrollbar); knobbles nPgNum, which scuppers the logic to
re-show this quickhelp only for new-pages. This is nastily painfull for people
with scroll-mice actually reading the text apparently.

This patch fixes the problem; may I commit (and to where) ?

--- sw/source/ui/uiview/viewport.cxx	22 Aug 2003 12:01:25 -0000	1.25.160.1
+++ sw/source/ui/uiview/viewport.cxx	7 Oct 2003 09:33:04 -0000
@@ -781,6 +781,7 @@ IMPL_LINK( SwView, ScrollHdl, SwScrollba
 		//Hier wieder auskommentieren wenn das mitscrollen nicht gewuenscht ist.
 		// JP 21.07.00: the end scrollhandler invalidate the FN_STAT_PAGE,
 		// 				so we dont must do it agin.
+ 		long nOldPge = nPgNum;
 		EndScrollHdl(pScrollbar);
 
 		Point aPos( aVisArea.TopLeft() );
@@ -801,7 +802,7 @@ IMPL_LINK( SwView, ScrollHdl, SwScrollba
 			USHORT nPageCnt = pWrtShell->GetPageCnt();
 			if( nPageCnt > 1 && Help::IsQuickHelpEnabled() )
 			{
-				if( !nPgNum || nPgNum != nPhNum )
+				if( !nOldPge || nOldPge != nPhNum )
 				{
 					Rectangle aRect;
 					aRect.Left() = pScrollbar->GetParent()->OutputToScreenPixel(
Comment 1 mci 2003-10-08 14:31:34 UTC
set target to OO1.1.1
reassigned to od
Comment 2 ulf.stroehler 2003-10-17 20:34:42 UTC
us->Michael: could you pls. explain what exactly is ment by "scrolling
tooltip pain" and attach a bugdoc that shows the misbehaviour.
BTW. what exactly does your patch? 
Thanks a lot.

us->od: please refer to issue 12927 for a possible bugdoc.
Comment 3 mmeeks 2003-10-24 12:15:00 UTC
The patch does something quite obvious - only throw up a new tooltip
with the page number / title in it when you hit a new page.

This stops a flickering / potentially massive tooltip permanantly
obscuring chunks of the text as you scroll.

To replicate, open any multi-page document, and use the scroll-mouse.
Comment 4 Oliver-Rainer Wittmann 2003-10-27 11:06:02 UTC
OD->Michael:
Thanks for your provided patch.
But only providing a new tooltip, if a new page is hit, isn't what the
tooltip should to. The tooltip should provide the actual page number
and the actual chapter.
I integrate your patch in local version of SRX645/m19 and tested it:
Using the mouse wheel to scroll doesn't change any behaviour - the
tooltip is still flickering. Using the scrollbar and scroll via mouse
by dragging and moving changes the behaviour as you described. Thus,
no new tooltip is shown, if a new chapter/subchapter is hit. That's
not, what we want.
Sorry, because of that, I will reject your provided patch.
I will submit a new issue, which will address the flickering problem.
Comment 5 Oliver-Rainer Wittmann 2003-10-27 11:07:55 UTC
closed
Comment 6 mmeeks 2003-10-27 14:03:02 UTC
If you are not using the code, you might want to cull the redundant
code in there - that _appears_ to have been designed to do precisely
what it does with my patch applied.
nPgNum is always zero when it gets to that branch so we simply was
cycles, and confuse the code flow to add this:

if (!nPgNum || nPgNum != nPhNum) - this is always going to be true.

Unfortunately, I don't have a scroll-wheel to benefit from the
flickering, which is a shame, the original bug was not down to
flickering, but the fact that the tooltip frequently obscures the text
you're trying to read.
Comment 7 Oliver-Rainer Wittmann 2003-10-27 15:03:58 UTC
OD->Michael:
Thanks for your feedback.
Yes, you're right: The code, you mentioned, seems to be redundant.
But, first, we have to provide the fix and then we can re-work the code.

The repaint errors of the normal text are already under investigation.
This defect seems to be platform-dependent - it doesn't occur on Windows.
Comment 8 utomo99 2003-12-22 05:37:07 UTC
Removing target from invalid Issue, to reduce the statistics.