> Was there any particular reason why tla--read-directory-name had this
> piece of code :
>
> (unless dir
> (setq dir default-directory))
> (unless default-dirname
> (setq default-dirname
> (if initial
> (concat dir initial)
> default-dir)))
>
I've just found this code is needed at least in GNU Emacs in CVS.
Next code follows the above code:
(if (featurep 'xemacs)
(read-directory-name prompt dir default-dirname mustmatch initial)
(read-file-name prompt dir default-dirname mustmatch initial)))
If dir and default-dirname are nil, read-file-name returns a *file* name,
not a directory name.
BTW, read-directory-name exists in GNU Emacs in CVS. So, I think we
should use:
(if (fboundp 'read-directory-name)
(read-directory-name prompt dir default-dirname mustmatch initial)
(read-file-name prompt dir default-dirname mustmatch initial))
instead of
(featurep 'xemacs)
emacs/lisp/ChangeLog.10:
2002-05-28 Kim F. Storm <[EMAIL PROTECTED]>
* files.el (read-directory-name): New function.
(cd): Use it instead of read-file-name.