John Levon wrote: > I'm not quite sure what behaviour has changed, though it might be the > work Alfredo did with top_y possibly ? Can you explain again what > behaviour changes you made for the graphics stuff ?
The change was pretty general (an safe): change all first_y as right value with top_y() and all first_y as left value with top_y(int). The only problem that can arise if there is some point where top_y() is called twice and it's assumed that it will return the same value (that can be not so, because some row height changed in previous rows in between). If you know in which piece of code this is happening, try to call top_y() only the first time and put this value on a variable. Note that there were also drawing problems before my change: Garst posted a testcase that came up pretty hosed with 1.3.x PS: if you really want to be sure, you can test the 'old behaviour' by simply: o add a global int first_y = 0 in text.C o change top_y(int) to only set this variable o change top_y() to only return this variable. Alfredo
