Is cell reference parsing wrong?

[leuski]

When cell_reference is parsing a string reference it looks like absolute flags are swapped. From the comments in the code:

Reference parameters absolute_column and absolute_row will be set to true if column part or row part are prefixed by a dollar-sign indicating they are absolute, otherwise false.

but in fact,

    if (column_string[0] == '$')
    {
        absolute_row = true;
        column_string = column_string.substr(1);
    }

    if (row_string[0] == '$')
    {
        absolute_column = true;
        row_string = row_string.substr(1);
    }

am I missing something?

[tfussell]

Yes, looks like something that broke in refactoring six years ago that hasn't been noticed since. There are unit tests for cell references, but only for both absolute column and row not one or the other. See https://github.com/tfussell/xlnt/blob/master/tests/cell/cell_test_suite.cpp#L667-L669 A fix with an extra test case or two is welcome otherwise I'll take care of it when I can.

[amiremohamadi]

@tfussell I'd like to work on it if no one is working on.
Just to make sure, if column part or row part are prefixed by a dollar-sign both of absolute_column and absolute_row must set to true?!

[tfussell]

@amiremohamadi Great! Contribution are always welcome. It's possible to have an absolute column with a relative row and vice versa. So there are four forms:

• $A$1 absolute column, absolute row
• $A1 absolute column, relative row
• A$1 relative column, absolute row
• A1 relative column, relative row

Official documentation is here https://support.office.com/en-us/article/switch-between-relative-absolute-and-mixed-references-dfec08cd-ae65-4f56-839e-5f0d8d0baca9

I would recommend turning these four into test cases in https://github.com/tfussell/xlnt/blob/master/tests/cell/cell_test_suite.cpp#L667-L669 and then writing the code that passes the tests (TDD).

[tfussell]

This should be resolved by #451.