Apache OpenOffice (AOO) Bugzilla – Issue 112076
Position offset of semi-transparent graph lines when printed
Last modified: 2017-05-20 11:42:11 UTC
How to reproduce in DEV300m80: - select a data series in a graph in Calc, and set transparency to non-zero value - print graph Result: Any data series with transparancy > 0% is positioned incorrectly in the graph. There is an offset in both the x and y positions. Tested with 3 different (networked) printers, using 3 different drivers: HPLJ 4350 and two "Develop" color copiers/printers. Used DEV300m80 from www.openoffice.org (on openSuSE 11.0, x86_64). Problem does not occurr in OpenOffice.org 3.2.0 on the same printers. Could this be a regression caused by the inclusion of CWS vcl106?
Created attachment 69783 [details] testcase with semi-transparent cosine graph
I see this also, testing OOO330_m6 on Fedora Linux 13; attaching screen captures comparing output from OOo 3.2.1 and OOO330_m6. Note: This is a regression compared to OOo 3.2.1. The sample problem affects thick semi-transparent lines printed from OOo Draw. Do we need a separate issue for that?
Created attachment 71488 [details] screen capture comparing "print to file" from 3.2.1 and 330_m6
Also, in the image I attached, you can see there is a difference in the chart frame rendering as well. It does not seem to be simply a matter of hairline rendering in the PS output, because the Postscript from the same chart, generated in OOo 3.2.1 and viewed at the same zoom, looks fine. Zooming in on the PS rendering does make some parts of the frame/axes visible, but not all.
Hi Iha, one for you. kla
My advice is not to expose end-users to this defect. This means: Target milestone:= OOo 3.3
Just used the test case with 3.3.0 beta 1: Problem is still there.
Back from vacation I found this issue assigned to me. I can confirm the problem also on windows. It is a regression compared to OOo 3.2.1. It seems to be a problem in the printing routine. Philipp, please have a look whether this can be fixed for 3.3 still. I consider this to be a serious bug. The printed charts show false data.
Created attachment 71871 [details] example showing the problem also with a draw ole in calc completely without charts
This does not seem to be an OLE object issue, however: if I copy the embedded drawing from iha's sample to a new Draw document, as normal drawing shapes, it shows the same problem when I print (to file) from Draw. The PDF export of the same drawing is fine.
fixed in CWS ooo33gsl10
Thanks, pl! I will test the BETA or RC release that merges this CWS. BTW: The new printing system fixes a lot of problems of my colleagues with OOo in one go. Excellent work, OOo crew!
@hi: please verify in CWS ooo33gsl10
Reassigned to OC
verified in internal build cws_ooo33gsl10
Testing "OOo-dev 3.3 330m10(Build:9533)" on Fedora Linux 13: Looks good: printing & print to file/Postscript render have the correct appearance for each test document. The middle two lines in the OLE sample render poorly, but I think that's a different problem related to embedded objects: the lines all render properly if I copy/paste them into a Draw document and print.
Verified as fixed in OOO330m12 (build: 9537) aka RC2 with testcase (attachment id=69783)