I'm using a plug-in that dumps out various getters of interfaces, as far I've needed them within the past 20 years, mostly on basis of simple iteration. Back in 2018 that came across ITableModel and CellContent, but I never covered ITableLayout or ITableGeometry because I did not need them. I also don't know how far they are covered in the programming guide.
Therefor we've accidentally come to talk about different interfaces, coming from entirely different models (table vs. text). My "not sure whether this is relevant".
Within the containing text model (my approach), kTableFrameBoss is an owned item, basically one per the 3 visual table fragments matching the 3 text columns. I never looked at the table frame though because I also ignored most of its interfaces, and in the 422k lines file produced from a minimum document I first came via table model to ICellContent - as mentioned.
Apparently my ICellContent::GetParcelFrameUID refers to a different object (the text colum kFrameItemBoss) while yours (still going thru ITableLayout::GetParcelFrameUID?) reaches the table frame.
If you look at ITableLayout, it has an own concept / nested class of parcel, frame etc. These are likely far more appropriate for your computations - I just don't cover them yet. E.g. see ITableLayout::Frame::GetParcelKey comment, looks like a bridge to reach the parcel enclosing the table frame. No separate types (TextParcelKey vs. TableParcelKey) though, because you can nest tables within tables ... See ITableFrame::GetFrameRef() - does it lead to my kFrameItemBoss or to the enclosing kMultiColumnItemBoss?
So with all these parcels around, nested frames (either table or text) and so forth, it is not trivial to even create test data to validate assumptions about what parcelkey is relative to what parcellist etc. Is that parcel key #0 relative here or there? What matrix is suddenly turned to non-ident by a cell stroke. Do you use the correct matrix, and all of them, or will it fall apart when you modify some cell padding (is there such a concept?). Add rotation somewhere, table within table, or disturb the flow, e.g. with anchored items or with text wrap on top of text cells, producing multiple parcels within a cell. Text wrap pushing into the outside text column producing multiple parcels there - how will the table frame stretch or split? Such an exhaustive effort will take several days just to produce test data. Can tables appear within footnotes and type-on-path?
Far beyond the scope of a forum post, where I don't know where in your larger context you're heading. Hit test, text drawing an own adornment, other decoration. Maybe the bounds and transforms in ITableFrame are your solution?
1 Correct answer
15
Replies
AdChoices