bug-ncurses
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: moving nested subwins


From: Bryan Christ
Subject: Re: moving nested subwins
Date: Thu, 07 Jun 2007 10:33:39 -0500
User-agent: Thunderbird 1.5.0.12 (X11/20070530)

I agree with Thomas. It is best to provide an extensions and not to mess with internal structures like struct _win_st. Although you could resolve the portability problems with something like #define NCURSES_MVWIN_FIXED, it requires a greater level of user awareness. For this reason, I think it would be best to offer an extension which recursively calculates offsets based on _parx/_pary and then corrects _begx and _begy.

Clive Nicolson wrote:
On Thu, 7 Jun 2007, Thomas Dickey wrote:

On Thu, 7 Jun 2007, Clive Nicolson wrote:

Oops I forgot to attach the code.

Hi,

here is some code which demonstrates how to fix the "menu subwin moving"
problem for mvwin. This code is nonrecursive and I think will not break user code which traverses the window parent tree and recalculates _begy
and _begx. It also fixes the "form subwin moving" problem!

mvderwin needs a somewhat slightly different fix.

Anyone want a patch?

not exactly. Let's go back to the top of the discussion. What's needed is a concise proposal for

(a) an identifiable mismatch between ncurses behavior that leads to a fix or

The problem is that [n]curses documents getbegyx thus:

"Like getyx, the getbegyx and getmaxyx macros store the current beginning coordinates and size of the specified window."

There is no statement about getbegyx returning the wrong values for a subwindow
if its parent window has been moved by mvwin, or have I missed it?


(b) an example of how to provide the functionality without modifying ncurses or

Others have done this, but it requires reaching into structures that should
be opaque.


(c) a new feature which could be used to provide the functionality, e.g., an extension

One could treat the fix as an extension by #define NCURSES_MVWIN_FIXED or some such mechanism.

I do not think my proposed fix will break any existing code, other than that
which uses values from getbegyx to somehow move a window and its subwindows
back to their positions before the parent window had been moved with mvwin.

My proposed fix provides the least surprise (in the values returned by getbegyx on a sub window) after a mvwin on its parent window.

As for compatibility with other implementations of curses, let us assume that ncurses is the leading implementation and that the others will follow
with a similar fix.

The only other fix would be to correct the documentation for getbegyx and
mvwin.

Clive


...




reply via email to

[Prev in Thread] Current Thread [Next in Thread]