Issue 13779 - anti-aliasing with full justifiction screen layout bug
Summary: anti-aliasing with full justifiction screen layout bug
Status: CLOSED FIXED
Alias: None
Product: gsl
Classification: Code
Component: code (show other issues)
Version: current
Hardware: All Linux, all
: P3 Trivial (vote)
Target Milestone: OOo 1.1 Beta2
Assignee: ulf.stroehler
QA Contact: issues@gsl
URL:
Keywords:
Depends on:
Blocks: 12697
  Show dependency tree
 
Reported: 2003-04-24 14:32 UTC by khendricks
Modified: 2003-05-20 10:46 UTC (History)
1 user (show)

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


Attachments
jpeg screen snapshot showing problem (see "estimate" and "average") (125.41 KB, image/jpeg)
2003-04-28 15:58 UTC, khendricks
no flags Details
small two line test case doucment (play with spacing in first line) (5.08 KB, application/octet-stream)
2003-04-28 16:02 UTC, khendricks
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description khendricks 2003-04-24 14:32:31 UTC
Hi,

It seems if I install stock ttf fonts (like Times New Roman - direct from 
the HP LaserJet official disk), and I make sure that text is full 
"justified" (both right and left and not ragged right
and that anti-aliasing is enabled,  I can literally make many
many leading letters of words appear to become disconnected from the
remaining letters of the word, but only under OOo 1.1 series (1.1 Beta 
and 644m10, etc) builds.

The problem only happens with BOTH full justification
and anti-aliasing.

It seems to occur when a "short" sentence needs to be spread out a
lot to become right justified and somehow the distance between the
letters in words are increased (along with the increased whitespace 
between words) resulting in the lead letter of a word becoming 
"detached" from the word itself (way far out in front).


I have exchanged e-mails with Herbert about this and I have
sent to him a small sample document and the fonts I have used
so that he might recreate this issue.

BTW, this issue was confirmed by another user on dev@porting
this morning.

I have included his confirmation below:

---snip---

And here was I thinking it was just a font kerning issue related to the
particular font I was using. I can confirm this on my Linux Beta 1.1
install. As you state, it only occurs with justified text. I see this
with the Helvetica font, for instance, and the font is fully
anti-aliased.

Alex.

---snip---

The bug in question only happens with 644/1.1 builds (it works fine in
OOO_STABLE_1 based builds.

In addition, when printed the problem does not appear in the
printed output.  The problem only appears on the screen.

I have tried to track it down using hints from Hamburg but
the pDXAry values look fine to me (but I must admit I am 
not sure how to interpret the debug output yet).

If someone can point me to the actual code that
does the justification part, I coould try to help track
down this issue further.

Thanks,

Kevin
Comment 1 christof.pintaske 2003-04-28 15:51:48 UTC
cp->hdu: please have a look at it
Comment 2 khendricks 2003-04-28 15:58:49 UTC
Created attachment 5936 [details]
jpeg screen snapshot showing problem (see "estimate" and "average")
Comment 3 khendricks 2003-04-28 16:02:24 UTC
Created attachment 5937 [details]
small two line test case doucment (play with spacing in first line)
Comment 4 khendricks 2003-04-28 16:07:33 UTC
Hi, 
 
I have attached a screen layout showing the problem (the same one I sent to Herbert). 
I have also attached a small text case (two line) Writer document. 
 
To see the problem simply load this file and dynamically add and remove letters and words 
and watch how the justification changes and how the leading character of some words 
begins to become detached. 
 
Hope this helps, 
 
Kevin 
 
ps. I would be happy to run any special tests or add debug output for you if it would help. 
 
Thanks, 
 
Kevin 
 
Comment 5 hdu@apache.org 2003-04-28 17:11:37 UTC
Seems to have slipped in when moving Sallayout::AdjustLayout() before 
glyph fallback step is completed to avoid crash on Win32's 
ScriptApplyLogicalWidth. 
 
Fix in progress. 
Comment 6 hdu@apache.org 2003-04-30 16:03:26 UTC
In vcl/source/gdi/sallayout.cxx Sallayout::GetNextGlyphs() the test at 
// stop when x-position unexpected 
was off by one. Fix is in revision 1.41.6.1. 
 
Thank you Kevin for catching this serious bug. 
Comment 7 hdu@apache.org 2003-05-02 10:04:04 UTC
to US for verification in BETA2REGRESSION03 
Comment 8 hdu@apache.org 2003-05-02 10:21:48 UTC
Second try to reassign. 
Comment 9 ulf.stroehler 2003-05-02 14:01:58 UTC
US: fix verified.
Comment 10 ulf.stroehler 2003-05-02 14:02:59 UTC
changing Status to verified.
Comment 11 thorsten.ziehm 2003-05-20 10:46:31 UTC
Verified in OOo1.1 Beta2.