I get this behavior with tabs. The second if block should not be highlighted

I get this behavior with tabs. The second if block should not be highlighted

This should be fixed now. (Please let me know if it not.) I used indent_state.stack to decide what indents were drawn, but I can see it doesn't work as I would expect in all cases (e.g. with the blank lines in the if statements like in the picture). To overcome this I have constructed a table drawn_indents alongside the whitespace_tbl that keeps track of exactly which indents have been drawn. This also simplifies a workaround with blank lines I used earlier.

I have added priority, show_start and show_end now. Note that show_end doesn't really make sense in the same way for the indentation guides as for scope, so the only thing I think makes sense here is to underline the line below the end of the current indent level if you enable it (and there is a line to underline).

I don't quite understand the need for drawn_indents, isn't it just the same information you already have in whitespace_tbl?

I don't quite understand the need for drawn_indents, isn't it just the same information you already have in whitespace_tbl?

The information in drawn_indents is already in whitespace_tbl, yes. I just thought it was easier to work with a list of purely the indents, whereas whitespace_tbl includes spaces and other stuff too. If you prefer, I can rewrite it to just use whitespace_tbl directly (you will just need to search backwards through whitespace_tbl for the indents).

drawn_indents just contains the (0-indexed) positions of TAB_START, TAB_START_SINGLE and INDENT in the whitespace_tbl (and it ignores TAB_FILL, TAB_END and SPACE), since these are the ones used for drawing the indentation guides.

👍 yeah that's what I mean. Let's just use whitespace_tbl. You can make a helper function to give you back the indents.
It shouldn't hurt performance, and it's easier to understand when there is only one data source.

👍 yeah that's what I mean. Let's just use whitespace_tbl. You can make a helper function to give you back the indents. It shouldn't hurt performance, and it's easier to understand when there is only one data source.

Since I only actually really need the last indent column and the number of indents I have made a helper function to return these given a whitespace table now.

Since the other pull request is almost ready, I will close this one.

Continued in #743 (adding a ref here because they were not linked). Thanks @Danielkonge for the great work.