Created attachment 201766 [details] correct and error I made the doc test1.docx in the Ms Word,but when I opened it in LibreOffice Writer，something is wrong,the pictures (for example 1/4) move up ,and not in line.But when I opened it in WPS or OnlyOffice ,it is OK. why?
I tried to do some bibisecting with this bug. It seems that the version 6.0.0.0.alpha1+ introduced a change that aligns the formula objects higher. In version 5.4.0.0.alpha1+ the formula objects are aligned lower (more correctly) although they are not aligned perfectly in this version either. I also noticed that in some newer versions of LibreOffice the formula objects are aligned higher and in some versions they are aligned lower but they are not aligned correctly in any of the versions I did bibisecting with. When doing bibisecting I marked the commits that align the formula objects higher as bad and the commits that align the formula objects lower (more correctly) as good. With this in mind I found out that the source hash of the first bad commit is 2a65ad5708d1013a28fe6c250701c8f385c2083f I used bibisecting repository bibisect-win32-6.0
I can confirm that the bug is present in the master build: Version: 26.2.0.0.alpha0+ (X86_64) / LibreOffice Community Build ID: 10850b9ea91807c4f8464e42fee7f33b7c389a77 CPU threads: 16; OS: Windows 11 X86_64 (build 26100); UI render: Skia/Vulkan; VCL: win Locale: fi-FI (fi_FI); UI: en-US Calc: CL threaded
I can confirm that the bug is present in the stable build: Version: 25.2.5.2 (X86_64) / LibreOffice Community Build ID: 03d19516eb2e1dd5d4ccd751a0d6f35f35e08022 CPU threads: 16; OS: Windows 11 X86_64 (10.0 build 26100); UI render: Skia/Vulkan; VCL: win Locale: fi-FI (fi_FI); UI: en-GB Calc: CL threaded
Did some bibisecting with bibisecting repository bibisect-win32-5.0. I might have found the first bad commit (source hash 3f1cbf94cb3ef743295eb7dcc8ed97e6252828ec). In version 5.0.0.0.alpha1+ the formula objects are positioned (almost) correctly. In version 5.1.0.0.alpha+1 the formula objects are positioned incorrectly (too high).
(In reply to Jussi Suominen from comment #4) > Did some bibisecting with bibisecting repository bibisect-win32-5.0. I might > have found the first bad commit (source hash > 3f1cbf94cb3ef743295eb7dcc8ed97e6252828ec). In version 5.0.0.0.alpha1+ the > formula objects are positioned (almost) correctly. In version > 5.1.0.0.alpha+1 the formula objects are positioned incorrectly (too high). Let's ask László what he thinks.