[upstream] Pressing the tab key in "overwrite" mode, deletes the char after point, and indents

[Dabian]

Binary package hint: openoffice.org

Pressing the tab key in "overwrite" mode, deletes the char after point, and indents. While this is logical and correct, it may not be what the user expects.

Steps to reproduce:

1. Write some text ("abcdefghijklmnopqrstuvwxyz").
2. Move the cursor to the beginning.
3. Press <insert> to activate overwrite mode, if not already enabled. press tab.

Actual behaviour:
The first letter (in this case the letter 'a') is replaced with the tabulator.

Expected behavor:
Not sure, one of the following I guess:

1. Insert a tab, like insert mode.
2. Skip to next tab stop without changing anything.
3. Skip to next tab stop, deleting all text between cursor position and next tab stop.

I guess the most logic would be 1 or 3.

[Chris Cheney (ccheney)]

For some reason hitting insert on hardy i386 doesn't change it to overwrite mode for me, but does on hardy amd64. For crimsun on IRC it works properly on i386 so I don't know what is wrong with my system.

[Chris Cheney (ccheney)]

Need to verify on upstream version once I can get the insert key working properly on my laptop.

[Ralph Janke (txwikinger)]

I can confirm this behaviour on Hardy with OpenOffice 2.4 on i386.

BTW. the "Insert" button does not work on my laptop either in order to toggle the mode, however, it is possible to toggle it by clicking on the statusbar on the field that shows the mode as insert/overwrite.

[Chris Cheney (ccheney)]

Confirmed on upstream's openoffice.org 3.0beta.

[Chris Cheney (ccheney)]

After discussing with the developer and further testing this is actually correct behavior.

Windows
--
Word 2007 - doesn't support overwrite mode at all
WordPad - supports overwrite and overwrites the character just like OOo

Linux
--
gedit - does what you want
vi - overwrites the character like WordPad and OOo.

Chris

[Dabian (dabian)]

I would like to know what is supposed to make this bug invalid?
That some proprietary program has the same behaviour doesn't make this bug invalid, IMHO.

Maybe it should be a feature request, but the behaviour has been confirmed. Maybe the problem is that Launchpad lacks a "wont fix" status?