Similar bug with text boxes (I updated the title for a more general formulation).

1. Create a text area with some text inside.
2. Select some text and drag-and-drop it outside the area. At this point LibO automatically created another text area top contain the pasted text.
3. Edit/Undo.

Result: the text is indeed restored inside the original text area but the new text area which was automatically created by the drag-and-drop operation was not removed (it should have been removed).

Confirmed with 4.1.4.2 on Debian x86.

please retest against 4.2.4.2
if bug persists please move it to mab4.2 list since 4.1.x is EOL

Right now I can test with 4.2.3.3 and the bug persists, but I now noticed if you press a second time Edit/Undo then the action is properly undone. Which means 1) there is a workaround! 2) the bug can be reformulated as: one extra action gets inserted in the history, so the user needs a double undo to get read of it.

I can't test right now with 4.2.4.2 but only 4.2.3.3 (still it's a 4.2.x) so I'll test again when the version you were asking for becomes available in my distribution.

in the meantime I move it to mab4.2 list
close it if it works in 4.2.4.2 (but I think that's unlikely)

Created attachment 110562 [details]
testcase

bug confirmed under Win8.1 64bit using LibO 4.3.4.1 and 4.5.0.0 alpha

load the attached test case.
move slide 1 after slide 3, then click edit/undo.

moving this to mab4.3 list since 4.2.x is end of life

works fine in LO 3.6.7.2 
bug present in 4.0.4.2
nece this is a 4.0.x brach regression, needs bibisecting

Bibisect results from 43all:
 3d8d766f50d0fd4ead4d4f7074a35bba465b365c is the first bad commit
commit 3d8d766f50d0fd4ead4d4f7074a35bba465b365c
Author: Bjoern Michaelsen <bjoern.michaelsen@canonical.com>
Date:   Tue Dec 11 08:13:31 2012 +0000

    source-hash-b67a51b40a4876f4bd97a2917103112006710b0c
    
    commit b67a51b40a4876f4bd97a2917103112006710b0c
    Author:     Markus Maier <maier@fs.ei.tum.de>
    AuthorDate: Thu Nov 29 16:10:23 2012 +0000
    Commit:     Caolán McNamara <caolanm@redhat.com>
    CommitDate: Thu Nov 29 17:35:25 2012 +0000
    
        add sort tab page .ui
    
        Change-Id: Ib5471b2a4cae45cf8aa32b438bac7f5cda35f71a

# bad: [423a84c4f7068853974887d98442bc2a2d0cc91b] source-hash-c15927f20d4727c3b8de68497b6949e72f9e6e9e
# good: [cf86b7f14a98d2d81a5cd93507acb35ff6775d8b] source-hash-85c6244b85b29c1d2bb9d89b62e9512dd65378b5
git bisect start 'latest' 'last35onmaster'
# bad: [34eab3946c46bb7273ba4ca395db9c4421dd232f] source-hash-e962805b31074d6b6a2ed0db6452769448337553
git bisect bad 34eab3946c46bb7273ba4ca395db9c4421dd232f
# good: [2e2d1aeff80dcbe390f6c3fbc54d6c8de81b0a0e] source-hash-64ab96cd15e52da88781e720d6f031dbcd0ba902
git bisect good 2e2d1aeff80dcbe390f6c3fbc54d6c8de81b0a0e
# bad: [9ac72638fdc1adf45f17c71f66baf9b20b606891] source-hash-1ed4210d8e6cbf1b25cd1eb8666be6578cc45845
git bisect bad 9ac72638fdc1adf45f17c71f66baf9b20b606891
# bad: [f0250a5564d9cdd406d788062cff81319971ccae] source-hash-db99a31c4e4e7c74d4c4bb7caa747a5752a32757
git bisect bad f0250a5564d9cdd406d788062cff81319971ccae
# good: [5809284e8a6f97fe3b5beb037b6b80eeb9fe1058] source-hash-246ffb108c7e1f762f8d497750ad2414b85b99ef
git bisect good 5809284e8a6f97fe3b5beb037b6b80eeb9fe1058
# bad: [4ef3a89e32ccba70e4b194ad97d110b88ce41c72] source-hash-8450a99c744e9005f19173e4df35d65640bcf5c4
git bisect bad 4ef3a89e32ccba70e4b194ad97d110b88ce41c72
# bad: [3d8d766f50d0fd4ead4d4f7074a35bba465b365c] source-hash-b67a51b40a4876f4bd97a2917103112006710b0c
git bisect bad 3d8d766f50d0fd4ead4d4f7074a35bba465b365c
# good: [f8b7f4b27eb8b3df4259a6d986607fd414f552af] source-hash-9d83ad0e99ab182506be99f7d6a2bec7f6fbe8c6
git bisect good f8b7f4b27eb8b3df4259a6d986607fd414f552af
# good: [e262cafc6f2ef071a248432b4d850e49d9e68c66] source-hash-c33019b36d613f951787ce9836e34d74bfbd6a1b
git bisect good e262cafc6f2ef071a248432b4d850e49d9e68c66
# first bad commit: [3d8d766f50d0fd4ead4d4f7074a35bba465b365c] source-hash-b67a51b40a4876f4bd97a2917103112006710b0c

It's not immediately obvious which commit in the above range could be at fault, and the regression is too old to conveniently build

There issue first appears in the range cbb599f657acc1953ef77d4f4fd895f75f9c7dad..22e54a9d0ee80b156ddb74a3d80330d48213ce9e, where there is some build breakage, but it seems probable that commit cbb599f657acc1953ef77d4f4fd895f75f9c7dad is the main issue here

commit 22e54a9d0ee80b156ddb74a3d80330d48213ce9e
Author: Andre Fische <andre.f.fischer Andre Fischerandre.f.fischer@oracle.com>
Date:   Fri Mar 11 15:22:17 2011 +0100

    impress211: #i110990# Fixed remaining problems with display ids and indices.
    
    Conflicts:
        sd/source/ui/dlg/present.cxx

commit 5c3614639cc448e5bdab6ed303de26a5ef9ce894
Author: Philipp Lohmann [pl] <Philipp.Lohmann@Oracle.COM>
Date:   Wed Mar 9 13:57:56 2011 +0100

    impress211: fix a warning

commit a46904855c32c306f42fce76735045d82013d993
Author: Philipp Lohmann [pl] <Philipp.Lohmann@Oracle.COM>
Date:   Wed Mar 9 13:50:30 2011 +0100

    impress211: fix a warning

commit 9177901fd6491be2937126839e5b96b1bb77ae7c
Author: Philipp Lohmann [pl] <Philipp.Lohmann@Oracle.COM>
Date:   Wed Mar 9 13:46:28 2011 +0100

    impress211: fix some warnings, types

commit 8bcf6ba5586818f4100113a6b685e891de84a5ea
Author: Andre Fische <andre.f.fischer Andre Fischerandre.f.fischer@oracle.com>
Date:   Wed Mar 9 11:16:39 2011 +0100

    impress211: #i110990# Fixed slide show spanning multiple displays on Windows.
    
    Conflicts:
        sd/source/ui/dlg/present.cxx
        sd/source/ui/slideshow/slideshow.cxx

commit cbb599f657acc1953ef77d4f4fd895f75f9c7dad
Author: Andre Fische <andre.f.fischer Andre Fischerandre.f.fischer@oracle.com>
Date:   Tue Mar 8 17:22:13 2011 +0100

    impress211: #i88880# Fixed dragging of slides from navigator to slide sorter.

("The issue..." - I need to typemo re carefully)

I fixed this with https://cgit.freedesktop.org/libreoffice/core/commit/?id=130f8eef234c6f44798d597814a0debc202fe8cc