File from older version. Heading 2 has background color set Open in 5.2.3 > There is no spacing Open in 5.3.0 daily master > There is spacing Obviously this causes all kind of wrapping issues. And I've seen no option yet to reset the spacing. A file created in 5.3. master doesn't have the issue. Will add test file.
Created attachment 129069 [details] test file from older version, having paragraph with background color
Regression introduced by: author Justin Luth <justin_luth@sil.org> 2016-10-31 08:59:09 (GMT) committer Miklos Vajna <vmiklos@collabora.co.uk> 2016-11-04 09:20:47 (GMT) commit 9130627e21dd7c52c5eee1acc4b71f86eb9f3118 (patch) tree 12faff6c8dee6626ee941d27e4da9016326ad7c5 parent c1ea561f6a01044357052789bbc6c8ec52061d41 (diff) tdf#41542 globally allow padding without borders: layout Styles with borders do not have a requirement for visible borders in order to apply padding in the ODF specs. This was only an implementation design, so remove that layout restriction. So, padding will be allowed now for any border (not just textframes/cells) and it will be allowed for any format (not only those marked with the compatibility setting AllowPaddingWithoutBorders - which is now redefined to indicate whether the UI allows creating borderless padding). The global UI implementation to allow creating these types of documents will come in 5.4. Allow the layout already in 5.3 to enhance compatibility between supported versions. Adding Cc: to Justin Luth
Padding *is* defined in the "header 2" style. Styles.xml contains fo:padding-left="0.15cm" fo:padding-right="0.4cm" fo:padding-top="0.4cm" fo:padding-bottom="0.199cm" fo:border="none". So actually not showing the padding was a bug. So do we consider this a regression that requires handling? The UI option to change the padding comes later in 5.4 (intentionally - see bug 41542). It is currently possible to reset the spacing by adding a border, applying, and then removing it. (OP's test document must have been made in a rather old version, since removing the border resets the spacing values to zero since before LO3.6).
Thanks for explaining (In reply to Justin L from comment #3) > So do we consider this a regression that requires handling? > [...] It is > currently possible to reset the spacing by adding a border, applying, and > then removing it. Since there is a work around and indeed, this comes from a really old version, I'm fine with not spending time on handling this exception. Now, to inform others about this, I guess a remark on the release notes for 5.3 is best?