bug#63949: 30.0.50; `vc-print-log´ does not erase buffer when called from *vc-change-log* buffer, at least for CVS logs

[Jens Schmidt]

On 2023-06-09  08:09, Eli Zaretskii wrote:

> What do you mean by "modifies the VC parent buffer to itself"?  Do 
> you mean that vc-print-log makes the *vc-change-buffer* be the VC 
> parent buffer?  If so, why is this incorrect?

Hope below recipe explains it.  The general idea is to start a VC
command first from the buffer of a VC-controlled file, then from the VC
result buffer, again from the VC result buffer, etc.

emacs -Q (from emacs-29)
<visit git-controlled file, say README>

;; current buffer equals README
C-x v l
;; current buffer equals *vc-change-log*
C-h v vc-parent-buffer
=> #<buffer README>

;; current buffer equals *vc-change-log*
C-x v l
C-h v vc-parent-buffer
=> #<buffer *vc-change-log*>

;; current buffer equals *vc-change-log*
C-x v l
;; git log of whole repository is displayed, not of README

I backed out commit d494833d47968fcd97ba549654a259d6fb6c2eee from
emacs-29 - both issues (the initial CVS-related one and the above one)
are gone after having done so.

Not all VC commands expose this problem.  `vc-annotate' seems to work
and even the VC parent buffer stays unchanged when invoking
`vc-annotate' repeatedly.  `vc-diff' gets the parent buffer wrong but
seems to work when invoked from buffer *vc-diff* repeatedly even if
parent buffer is wrong.