Issue 52885 - JPEG-compressing PDF export shows tif as negatives
Summary: JPEG-compressing PDF export shows tif as negatives
Status: CLOSED DUPLICATE of issue 47395
Alias: None
Product: gsl
Classification: Code
Component: code (show other issues)
Version: 680m122
Hardware: PC Windows XP
: P3 Trivial (vote)
Target Milestone: OOo 2.0
Assignee: philipp.lohmann
QA Contact: issues@gsl
URL:
Keywords: regression
Depends on: 47395
Blocks:
  Show dependency tree
 
Reported: 2005-08-04 15:17 UTC by Rainer Bielefeld
Modified: 2005-08-08 10:44 UTC (History)
1 user (show)

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


Attachments
screenshot showing the effect (292.31 KB, application/vnd.oasis.opendocument.graphics)
2005-08-04 15:21 UTC, Rainer Bielefeld
no flags Details
Testkit for check with CWS vcl42 (225.22 KB, application/x-compressed)
2005-08-06 16:32 UTC, Rainer Bielefeld
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description Rainer Bielefeld 2005-08-04 15:17:33 UTC
I checked with 2.0 (1.9.m122) German version WIN XP: [680m122(Build8941)] and
found the PDF-Export completely messed up, pls. see attached screenshots
comparing results of 680m113 (ok) and 680m122 (messed up).

This issue may be a regression of issue 47395

Something concerning the DRAW document from which I exported:
I use the document to add changes and additions to an old circuit diagram. For
that I scanned the old diagram and inserted the .tif drawings by link into pages
of my draw document to Layer "Layout". That layer will be locked for further work.

All my changes will be in a new Layer "Änderungen" (= "changes"), for example
texts (as "Modul:01" and "-4N2").

The missing Text "$22: 000" ist a text field like  "Modul:01" 

Many other elements of the drawing are missing in the PDF export, sometimes I
even see a border where the element should have been, sometimes not. Those
2missing" Elements (at least the text Elemnts) seem to be in the PDF (I can find
them with PDF text search", but they are invisible, because black as the inverse
.tif area. "Modul:01" is visible, because it is written on a white background.

That all seems to say that this really is a regression from issue issue 47395,
but I am a little worried concerning the small PDF size even if I try
"uncompressed", but may be that that is caused by the many "black line on black
background", what must not be shown in the PDF grafic?

If you believe it might eb useful, I can supply a drawing with only few pages to
demonstrate the most important problems.

Please decide whether this issue should remain or be closed DUP.
Comment 1 Rainer Bielefeld 2005-08-04 15:21:36 UTC
Created attachment 28495 [details]
screenshot showing the effect
Comment 2 Rainer Bielefeld 2005-08-04 15:25:24 UTC
T believe that it is related to issue 47395 because of comments in issue 52851
Comment 3 Rainer Bielefeld 2005-08-04 15:29:50 UTC
It's nothing with the DRAQ-document itself, PDF export from the same document
with 680m113 works without problems.
Comment 4 philipp.lohmann 2005-08-04 16:02:04 UTC
It's not a regression since the fix for issue 47395 has not been integrated. If
that tiff is 1 bit and you used "lossless compression" this seems like a
duplicate to 47395.
Comment 5 Rainer Bielefeld 2005-08-04 16:30:22 UTC
The effect is always visible, with "looseless compression"  and also with "jpeg
compression".

Does "tiff is 1 bit" mean "black / white"? Yes, the tifs are "monochrome".
Comment 6 Rainer Bielefeld 2005-08-04 17:21:53 UTC
I checked with 2.0 (1.9.m118) English version WIN XP: [680m118(Build8936)] and
also saw the problem (m113 was ok!).
Comment 7 philipp.lohmann 2005-08-04 17:39:58 UTC
As long as the graphic is 1 bit monochrome it does not matter what compression
you use (1bit images cannot be JPEG compressed); sorry i was unclear about that.

Just one question remains: is it the same TIFF file that worked in 113 and does
not in 118 ? Or might two different TIFF files in use where the meaning of 1/0
for black and white pixels is simply reversed ?
Comment 8 Rainer Bielefeld 2005-08-04 19:20:39 UTC
I did all my tests with the same DRAW document containing the all the same tiff-
images.
Comment 9 Rainer Bielefeld 2005-08-04 20:21:05 UTC
Am a little sluggish today, but I begin to understand. In 680m113 we had the
same problem, but it occured only for "looseless compression".

Later versions (starting for me with 680m118) now show the same bug also for
"jpeg-compression".
Comment 10 Rainer Bielefeld 2005-08-04 20:37:50 UTC
Amended summary to current standard of knowledge
Comment 11 philipp.lohmann 2005-08-05 09:35:19 UTC
Well, the 1 bit tiff problem doesn't happen anymore in CWS vcl42. For
verification that this issue is actually fixed with issue 47395 i'd like to try
test one of your sample documents. Could you attach one  ?
Comment 12 Rainer Bielefeld 2005-08-06 16:32:52 UTC
Created attachment 28536 [details]
Testkit for  check with  CWS vcl42
Comment 13 Rainer Bielefeld 2005-08-06 16:35:06 UTC
'testkit.zip' contains a drawing with some (linked) tif-files. With that example
I see the "negative-problem" with 2.0 (1.9.m122) German version WIN XP:
[680m122(Build8941)] in looseless _and_ JPEG- compression.
  
Comment 14 philipp.lohmann 2005-08-08 10:44:11 UTC
Ok, this is reproducible in m122 but not in CWS vcl42 (based on m122) anymore.
That makes it a duplicate to issue 47395 after all.

*** This issue has been marked as a duplicate of 47395 ***
Comment 15 philipp.lohmann 2005-08-08 10:44:40 UTC
closing duplicate