Download it now!
Bug 120760 - FILEOPEN Z-order of objects in attached DOCX is wrong
Summary: FILEOPEN Z-order of objects in attached DOCX is wrong
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
(earliest affected) release
Hardware: All All
: medium normal
Assignee: Szabolcs Toth
Keywords: bibisected, filter:docx, regression
Depends on:
Blocks: DOCX-Header-Footer WPS-Support
  Show dependency treegraph
Reported: 2018-10-21 23:50 UTC by Aron Budea
Modified: 2020-06-23 07:49 UTC (History)
3 users (show)

See Also:
Crash report or crash signature:

Sample DOCX (39.47 KB, application/vnd.openxmlformats-officedocument.wordprocessingml.document)
2018-10-21 23:50 UTC, Aron Budea
Comparison screenshot (50.20 KB, image/png)
2018-10-21 23:50 UTC, Aron Budea

Note You need to log in before you can comment on or make changes to this bug.
Description Aron Budea 2018-10-21 23:50:12 UTC
Created attachment 145883 [details]
Sample DOCX

The attached DOCX (created in Word) contains two images and a text box in the header.
One image is the page background (light blue), and should be behind the others (a red box and text box with green background).

=> It's shown in front instead (except the text in the text box). It's a Z-order bug, because if the image Picture 2 (the background) is deleted in the Navigator, the others are shown.

Observed using LO (2018-10-09, 425af6845ebe066c950b0b63f50563e067485f3e) & / Windows 7.
In the text box is completely in front of the background, that perhaps could be bibisected.
In the red image is in front of the background.
Comment 1 Aron Budea 2018-10-21 23:50:47 UTC
Created attachment 145884 [details]
Comparison screenshot
Comment 2 Xisco Faulí 2018-10-22 07:26:43 UTC
Hello Aron,
I've bisected the problem with the green textbox and it points to 57450afb768c085df0ba2344aa94b5f843060178, which is already mentioned in other bugs ->
Regarding the red shape, I think it's inherit from OOo
I let you decide what to do with this issue...
Comment 3 Aron Budea 2018-10-22 21:38:04 UTC
Thanks for the bisect, Xisco! Since wps support was new development, these days I'm inclined to not regard that as a regression, but as implementationError.

The issue with Z-order of the red image was introduced in the following range (bibisected using repo bibisect-43all:

I'm inclined to think either of the following commits from Cedric are responsible:
author		Cédric Bosdonnat <>	2012-07-31 14:46:59 +0200
committer	Cédric Bosdonnat <>	2012-07-31 15:22:23 +0200
n#772094: pictures was eaten by document default LR settings
author		Cédric Bosdonnat <>	2012-07-31 15:19:23 +0200
committer	Cédric Bosdonnat <>	2012-07-31 15:22:23 +0200

n#772094: writerfilter, pictures anchored in header/footer won't be opaque