set-buffer-multibyte problems (from Bug #495089)

[Uday Reddy]

This is left over from Bug #495089
The problem here was caused by the call to set-buffer-multibyte, which is a questionable function. There are 13 remaining calls to this function in the current version. I will keep this bug report alive until I have looked at all the 13.

[Uday Reddy (reddyuday)]

Ulrich, can you take a lead on this please? I can look at any instances that you flag up for me. -- Uday

[Ulrich Müller (ulm)]

I believe that the following calls are o.k., since they are either done for an empty buffer, or immediately followed by a call to erase-buffer:
   vm-edit.el:73: (set-buffer-multibyte nil))
   vm-mime.el:2733: (set-buffer-multibyte nil))
   vm-mime.el:3112: (set-buffer-multibyte nil))
   vm-mime.el:5237: (set-buffer-multibyte nil))
   vm-pine.el:1154: (set-buffer-multibyte nil))
   vm-window.el:126: (set-buffer-multibyte nil))
   vm-window.el:145: (set-buffer-multibyte nil))

The following is redundant and should be removed (vm-make-work-buffer is guaranteed to create a single-byte buffer):
   vm-mime.el:4609: (set-buffer-multibyte nil))

Still to be checked:
   vm.el:134: (set-buffer-multibyte nil))
   vm-mime.el:321: (and vm-fsfemacs-p (set-buffer-multibyte t))
   vm-mime.el:1956: (and vm-fsfemacs-mule-p (set-buffer-multibyte t))
   vm-reply.el:980: (set-buffer-multibyte t)))
   vm-save.el:65: (set-buffer-multibyte nil))

[Ulrich Müller (ulm)]

Status update:

I've removed the following call to set-buffer-multibyte in the 8.1.x branch:
   vm-mime.el:4609: (set-buffer-multibyte nil))

This call is o.k., it is followed by erase-buffer:
   vm-save.el:65: (set-buffer-multibyte nil))

Leaves us with the following four, line numbers are for revision 746 in the 8.1.x branch:
   vm.el:139: (set-buffer-multibyte nil))
   vm-mime.el:321: (and vm-fsfemacs-p (set-buffer-multibyte t))
   vm-mime.el:1948: (and vm-fsfemacs-mule-p (set-buffer-multibyte t))
   vm-reply.el:981: (set-buffer-multibyte t)))