<<< Date Index >>>     <<< Thread Index >>>

Re: Searching in the index is not documented



Guten Tag, Michael!

On Thu, Aug 17, 2006 at 12:28:38PM +0200, Michael Tatge wrote:
> * On Thu, Aug 17, 2006 Alan Mackenzie (acm@xxxxxx) muttered:
> > This manual, fine though it otherwise be, does not describe how to do
> > searching within the message index.

> It does you just have to know where to look :)

This is precisely my point.  I started using mutt at the end of May 2006.
As a beginner, I didn't know where to look.  Despite searching long and
hard in the manual for the answers to these questions, I didn't find
them.  (Just as a matter of interest, I'm an Emacs hacker, and have a
great deal of experience in revising one of the Emacs manuals.)

Hence, I'm suggesting to the Mutt team, in a respectful fashion, that it
could be helpful to enhance the manual.

> > If I type "/foo":

> simple_search

>    Type: string
>    Default: "~f %s | ~s %s"

>    Specifies how Mutt should expand a simple search into a real search
>    pattern. A simple search is one that does not contain any of the ~
>    operators. See ``[1182]patterns'' for more information on search
>    patterns.

>    For example, if you simply type joe at a search or limit prompt, Mutt
>    will automatically expand it to the value specified by this variable.
>    For the default value it would be:

>    ~f joe | ~s joe

OK.  This answers my point (i), I think:  Namely, when I enter "/foo" in
the index display, mutt searches for the string in the From: and Subject:
headers.  Thanks!  I would have found it useful to have read this
explicitly in the manual 2 months ago.

Incidentally, what is the significance of the spaces in "~f joe | ~s
joe"?  Does "~fjoe" mean the same as "~f joe"?  What about "~f   joe"?
The manual seems to indicate a single space is needed, but is
frustratingly vague about it.

Incidentally(2), can I search for a literal "~f" in the Subject: header
by a typing something like: "/\~f"?

> > (ii) Is "foo" interpreted literally as text or as a regular expression?

> Let's check the "Patterns" section:
> ~f EXPR       - messages originating from EXPR
> ~s EXPR       - messages having EXPR in the ``Subject'' field.
> "Where EXPR is a regular expression."

Yes, I've got that now.  But I had to discover it by experiment back in
May, since at the time I didn't have enough background knowledge of mutt
to know that the word "Patterns" was pertinent to searching.

> > (iii) If there are several instances of "foo", in what order does mutt
> >   find them?  I would guess that it would use the order the mails are
> >   currently sorted in.

> correct.

Why do I have to guess?

> > (iv) Can I somehow search my mailbox for a pattern?  (e.g. "~fAlan")?

> Of course. What good are patterns when you cannot use them to search?

They're good for tagging (with `T'), or for restricting the messages
displayed (with `l').  I didn't learn that they could be also used for
searching (with `/' and `M-/' (excuse the Emacs notation ;-)) until
Tuesday of this week, when I tried it out.  I would have learnt this much
earlier if the manual had been more explicit.

> HTH,

Yes, it's helped a lot.  Thanks!

> Michael

-- 
Alan Mackenzie (Munich, Germany).