Re: spam-get-article-as-filename needed?

[Kevin Greiner <kgreiner@xpediantsolutions.com>]

Ted Zlatanov <tzz@lifelogs.com> writes:

> On Mon, 24 Feb 2003, kgreiner@xpediantsolutions.com wrote:
>> I was looking at the compilation logs only to see that spam.el
>> generated a couple of warning messages concerning
>> nnml-current-directory and nnml-possibly-change-directory.
>
> Strange, I was not seeing those warnings.

If you compile in the same emacs as you are running gnus, you won't
see these as gnus will have already loaded nnml.  I only see this sort
of message when I compile in a new emacs that has never run gnus.

>> After checking a little further, I found that
>>   1) The only dependencies to nnml are in
>>      spam-get-article-as-filename.
>>   2) spam-get-article-as-filename is functionally equivalent to
>>      nnml-article-to-file except that the later appears to handle
>>      compressed files (see nnml-use-compressed-files).
>>   3) I can't find any calls to spam-get-article-as-filename.
>> 
>> I don't want to touch spam.el since I'm not at all familiar with
>> that code.  Still, I'd like to see the compilation warnings
>> resolved.
>> 
>> It seems to me that the options are:
>> 1) Add (require 'nnml) to spam.el
>> 2) Delete spam-get-article-as-filename
>> 3) Rewrite spam-get-article-as-filename to implicitly depend on
>>    nnml-article-as-file.  For example,
>>       (defun spam-get-article-as-filename (article)
>>         (funcall 'nnml-article-to-file article))
>
> The original intent was (and still is) to allow training a spam/ham
> backend directly from the files, rather than through the less
> efficient parameter passing.  It's not useful yet, however, so I
> commented the code out for the time being.  It will come back when I
> put a bunch of other things in place, and I'll keep in mind your issue
> with nnml symbols.
>

Thanks.  That's what I hoped to hear.

Kevin