https://bugs.documentfoundation.org/show_bug.cgi?id=174221
--- Comment #7 from Oleg <[email protected]> --- (In reply to Saburo from comment #6) > I am not very familiar with .doc files, so I will not be changing the status. > > adding CC: Justin Luth > Please, can you take a look? > Fix commit 565912fa for tdf#117217 > Could the fact that there are two lines for `NS_sprm::TDefTableShdRaw::val` > be the cause? > /core/sw/source/filter/ww8/wrtww8.cxx Thank you, Saburo, that is exactly it. Confirmed in the code: commit 565912fa (tdf#117217) defines sal_uInt32 const aSprmIds[] { NS_sprm::TDefTableShd::val, NS_sprm::TDefTableShdRaw::val, NS_sprm::TDefTableShdRaw::val, <-- listed twice NS_sprm::TDefTableShd2nd::val, NS_sprm::TDefTableShdRaw2nd::val, NS_sprm::TDefTableShd3rd::val, NS_sprm::TDefTableShdRaw3rd::val }; The pattern of the list (Shd, ShdRaw, Shd2nd, ShdRaw2nd, Shd3rd, ShdRaw3rd) suggests that the second TDefTableShdRaw line is an accidental duplicate. As far as I can tell from the diff, for tables with up to 22 columns the loop writes TDefTableShdRaw twice with identical content, which matches the files I attached: 0xD612 (TDefTableShd) once and 0xD670 (TDefTableShdRaw) twice per row. So the sprm in question is sprmTDefTableShdRaw (0xD670), more precisely than my initial title says. This also fits the regression range: 6.4.7.2 (before the commit) writes it once; 24.2 and 26.8.1.1 write it twice (the commit went into 7.1.0). I have not tested versions between 6.4.7.2 and 24.2 myself. Suggested fix: remove the duplicate NS_sprm::TDefTableShdRaw::val entry in aSprmIds in WW8AttributeOutput::TableBackgrounds() (sw/source/filter/ww8/wrtww8.cxx). A unit test could check that no table-row property set contains the same sprm twice. Justin, could you please take a look? If you agree, "Regression introduced by: 565912fa" can be set. -- You are receiving this mail because: You are the assignee for the bug.
