possible emacs bug, need help with repro

[Le Wang]

Hi all,

I discovered a weird issue with Emacs that I reproed with NTEmacs 23.2.1 and CVS build from 20110112.

I need someone to follow some of these steps to see if they can repro the same thing on another platform before I file a bug report.  You need the "rebox2.el" attached.

1. emacs -Q rebox2_bug.el
2. M-x eval-current-buffer
3. C-x b<enter> ;; switch to "*scratch*"
4. M-x rebox-mode
5. M-<
6. SPC,SPC ;; space twice
7. C-aC-aC-aC-a ;; notice that the point is stuck at 6
8. C-xb<enter> ;; switch to "rebox2_bug.el
9. M-g g 1714
10. C-d ;; uncomment the hack workaround the bug
11. C-M-x ;; reevaluated the function 
12. C-x b<enter> ;; switch to "*scratch*"
13. C-aC-aC-aC-a ;; notice that the cursor switches between 1 and 6

So step 10 fixed the issue I was debugging with rebox2, however, it does nothing except refer to the marker "marked-point" in a format call.

(format "%s" marked-point)

How can this line change Emacs' behaviour?

If you try the steps, please let me know if you can confirm my observations.  Thanks.

--
Le

rebox2_bug.el
Description: Binary data