https://bugs.documentfoundation.org/show_bug.cgi?id=122098

Justin L <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[email protected]
            Version|5.2 all versions            |4.2 all versions

--- Comment #9 from Justin L <[email protected]> ---
repro 7.4+ and confirmed the identified commit in comment 6 with
bibisect-42max.

This is in some way just a cell format issue. Prior to the regression, this was
formatted as a Boolean Value, but afterwards as a Number-General.

There is some kind of automatic formatting going on here though, because we
don't have any formatting defined in the xlsx. Specifically, there is no
<numFmts count="2">
  <numFmt numFmtId="164" formatCode="GENERAL"/>
  <numFmt numFmtId="165" formatCode=""TRUE";"TRUE";"FALSE""/>
</numFmts>

and the cell references style s=0 (which seems to be a special
nothing-default?)

Interestingly, if I add a new function [=Exact(b6,c6)], then it automatically
uses a boolean value that appears to be directly formatted (because clear
direct formatting results in General format - aka "1"). However, on a
round-trip that formatting is lost. But if I explicitly set the cell as Boolean
Value, then it is retained.
Exporting this broke in LO 4.3, with author Kohei Yoshida on 2014-03-11
19:09:41
commit  69ecdad805281b2cb6ec2437da18daa19576deae
fdo#74747: Correctly inspect formula result value for xlsx export.


-------------------------------------------------------------------------------
INTERESTING: if I save as ODS, then (even though it originally only showed 1,0)
it imports as true/false. In fact the opposite is true. Clearing the direct
format settings in A1 shows "1", but on a round-trip it again shows TRUE.

This comes from  sc/source/filter/xml/xmlcelli.cxx  SetFormulaCell()
    if (bMayForceNumberformat)
        // Re-calculate to get number format only when style is not set.
        pFCell->SetNeedNumberFormat(!mbHasStyle);

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to