make it so escape make the introspection window in the notebook disappear

[williamstein]

this was the most commonly requested features during Math 480B

Component: notebook

Issue created by migration from https://trac.sagemath.org/ticket/5644

[qed777]

See comment.

[qed777]

comment:1

Attachment: trac_5644_esc_kills_introspect.patch.gz

The attached patch seems to work in Firefox 3 without breaking other functions. I'm not sure about the purpose of

            if(browser_op) { focus_delay(id,true); }

Although I haven't copied this to the new conditional clause, the patch works in Opera 9, too.

By the way, I think the halt_introspection(); statement just before the new code is unnecessary, since handle_replacement_controls() calls it before returning true. But the patch leaves this alone.

[qed777]

comment:2

To clarify: This patch works for a "TAB" introspection window but not for a "SHIFT-ENTER" introspection window. The latter is placed in cell_output_* instead of introspect_div_*, which is more difficult to make disappear, at least for me.

Because of this, I've changed this ticket's status to "needs work." The TAB-only patch may be enough for some users.

[qed777]

comment:3

On "SHIFT-ENTER" introspection: If you're feeling adventurous, replace the new lines in the patch with these:

    } else if((introspect_id == id) && introspection_loaded) {
        if(key_close_helper(e)) {
            halt_introspection();
            return false; //otherwise, keep going
        }
    } else {
        output_html = get_element('cell_output_html_'+id);
        pre0 = output_html.getElementsByTagName('pre')[0];
        if (pre0 && (pre0.getAttribute('class') === 'introspection') && key_close_helper(e)) {
            output_html.removeChild(pre0);
            return false; //otherwise, keep going
        }

Note: This doesn't tell the server about any changes! I don't know exactly how this plays out in practice.

[qed777]

comment:4

Ultimately, perhaps it's better to "unify" the treatment of introspection windows. That is, should the notebook place both TAB and SHIFT-ENTER results in a cell's dedicated introspect_div_*? Then binding ESC to halt_introspection() should be enough. This may also simplify related code (e.g., see #6001).

At first glance, which is all I can muster right now, notebook treats TABS and SHIFT-ENTER differently enough that this is not a trivial change.