Word Column header reading issue.

[nvaccessAuto]

Reported by vgjh2005 on 2014-11-11 15:06
Hi:
I've recieved a file that set the first column that there are two big cells merged by two small cells to be the column header. NVDA cannot read the correct header in the forth column the second line. Please open this file, then switch to the second line by pressing ctrl + alt +down arror, then press control + alt + right arror to jump to the forth cell.

[nvaccessAuto]

Attachment sample for office word.docx added by vgjh2005 on 2014-11-11 15:07
Description:
Update:
File added from Trac
sample for office word.docx.zip

[nvaccessAuto]

Comment 2 by mdcurran on 2014-11-19 07:03
NVDA throws exceptions when trying to get the headers for the 4th cell on the row, but the more important issue is actually that the headers for cells 2 and 3 are also wrong.
Technical:
MS Word changes the numbering of its cells based on merges, and as far as I can see, does not expose column span information.
So in this example, according to MS Word's object model, on row 1, the first two cells have been merged to form one cell (column 1). Therefore column 3 has now become column 2, which is not right as far as NVDA's header calculations go.
The exceptions at column 4 is simply that the above row only contains 3 columns now, not 4.
I will investigate this further.

[LeonarddeR]

@dkager: Looks this is the issue we were talking about privately. Possibly related to #6233

[dkager]

I wonder what exactly the issue is. Is it misreading colspans as in #6233 or is it also that you get exceptions?

[Adriani90]

@dkager I am testing with NVDA 2019.1.1 in Word 2016 / 365, UIA is disabled. NVDA missreads the column headers. In column 3, NVDA reads the header for column 4 through 5 and in column 2 it reads the header for column 3.

@michaelDCurran it seems enabling UIA solves this issue.