When you have a cell set to wrap the text into multiple lines, adding text exceeding the last line fitting in the cell's height results in the row heigh increasing, to fit the extra text. But - if you then shorten the text, so that it doesn't need the extra cell height - the row height doesn't go back down. The row height does get readjusted if you disable wrapping for the cell and re-enable it. I would like the row height to get recalculated not just when exceeding the visible rows, but when removing text as well. I haven't noticed an option to enable this automatic downwards-adjustment; assuming it indeed does not exist, there are actually two requests here: * Support for this behavior as an option, and * Making this behavior the default I would like both, but just the first is already good, provided I can choose to make it the default in new documents. I remember this behavior from ages ago - probably back from OOo.
(In reply to Eyal Rozenberg from comment #0) > I remember this behavior from ages ago - probably back from OOo. This used to be the behavior up until LO 7.1 (included). I know that in LO 7.2.7.2 the behavior is already as it is in current versions (i.e. the height of the cell does not shrink back when deleting part of the text on a cell that has the Wrap Text property). Please be aware that there are other regressions reported regarding the "Automatic Wrap Text" property, with multiple dupes among those regression reports.
(In reply to ady from comment #1) > This used to be the behavior up until LO 7.1 (included). I guess I have "anti-rosy glasses" for thinking about the past then... Anyway, yeah, there are a lot of row-height-related regressions - but I haven't found one about this specific issue.
(In reply to Eyal Rozenberg from comment #2) > (In reply to ady from comment #1) > > This used to be the behavior up until LO 7.1 (included). I realize I should clarify, JIC. The requested behavior in this report used to be the normal behavior until LO 7.1 (included). The behavior changed after that, probably somewhere LO 7.2 – it is already not shrinking on LO 7.2.7.2. IDK whether the change was/is intentional. As for other regressions, I meant regarding the Automatic Wrap Text property. See for example bugs 159351, 159834, 163099, 163150, all dupes of bug 159690 (and also some topic(s) in the Ask site).
reproducible with Version: 25.2.0.0.alpha0+ (X86_64) / LibreOffice Community Build ID: 4787fd4fc86230893a6da309f45964116b3a67df CPU threads: 4; OS: Linux 6.8; UI render: default; VCL: gtk3 Locale: cs-CZ (cs_CZ.UTF-8); UI: en-US Calc: threaded but not in Version: 6.2.0.0.alpha1+ Build ID: a20a2d7e0d28658f2d9089da076961a599833a28
This seems to have begun at the below commit in bibisect repository/OS linux-64-7.2. Tor is not active anymore, adding CC to Michael. 649a486aded9c34e0ed3557f4b0b1bc886e4c9f6 is the first bad commit commit 649a486aded9c34e0ed3557f4b0b1bc886e4c9f6 Author: Jenkins Build User <tdf@pollux.tdf> Date: Tue Feb 2 13:55:10 2021 +0100 source e763e13873adfe3c6abfa4c2dfd3ac3847e2d494 110285: Don't bother shrinking row height when changing just one row interactively | https://gerrit.libreoffice.org/c/core/+/110285 Change is intended because of collaborative editing: 110696: Don't shrink row height when deleting cell contents interactively either | https://gerrit.libreoffice.org/c/core/+/110696
FWIW, automatic dynamic row height is involved in more than a couple of reports, including some import/export (from/to xls(x)) reports. I don't know whether Automatic Wrap Text always implies automatic dynamic row height (in both directions).