Nick Dokos <ndokos@gmail.com> writes:

> Rainer M Krug <Rainer@krugs.de> writes:
>
>> Vitalie Spinu <spinuvit@gmail.com> writes:
>>
>>>  >> Rainer M Krug <r.m.krug@gmail.com>
>>>  >> on Fri, 7 Jun 2013 17:40:53 +0200 wrote:
>>>
>>>  > On Friday, June 7, 2013, Vitalie Spinu wrote:
>>>  >     All your examples are placed in fundamental mode. The comments are
>>>  >     treated by org and thus are correct, local variables are inserted
>>>  >     according to the major mode.
>>>
>>>  > The question is why - all .R files are automatically in r mode when I open them
>>>  > and all other R files tangle fine.
>>>
>>> Because they are placed automatically in R mode, your files are in
>>> fundamental mode.
>>
>> I think I get it now: my files are interpreted by emacs as fundamental
>> files. But when I visit them in a buffer, the buffer is interpreted as
>> an R buffer.
>>
>
> Almost. Files are *opened* in some mode: if no other mode is
> found, they are opened in fundamental mode - by themselves, they are not
> "fundamental files" or "R-mode files" or ... That's an interpretation
> that emacs superimposes on them when it visits them. To do that, it uses
> information in auto-mode-alist, or in any mode specification in the file
> itself (using file-local variables embedded in comments).
>
> So when you open a file "foo", it will be opened in fundamental mode
> (absent some other specification in the auto-mode-alist). When you add
> contents, the mode does not change. If you save it as file "foo" and
> then reopen it, it will still be in fundamental mode.  But if you save
> it as "foo.org" and then reopen it, the auto-mode-alist will tell emacs
> to change the mode to org-mode (assuming that the .org suffix has been
> set up correctly in the auto-mode-alist).  It does not matter what the
> contents were: it could be a C program e.g.  and it would still be
> opened in org mode.

Thanks for this very clear description of how this happened. I think I
understand what is going here.

And now I see why the block NAMESPACE file uses the R comment character,
while the others don't: NAMESPACE is in the auto-mode-alist defined as
an R file.
Adding DESCRIPTION as an R file would not be useful, as it is not an R
file - it follows the Debian Control File rules. I dont think there is a
mode for these?

But wouldn't this mechanism be an argument for a header argument to
enable adding of the mode of the file as tangled?
Always adding a local variable to the tangled file might cause problems
with certain formats which do not support comments.

Cheers,

Rainer

-- Rainer M. Krug

email: RMKrug<at>gmail<dot>com