9fans archive / 2006 / 05 / 248    prev next


    search terms are split using tokenize    
    from:regexp searches From: lines    
    subject:regexp searches Subject: lines    
    before:yyyy[/mm[/dd]] and after:yyyy[/mm[/dd]] specify date range    

    powered by grep(1)    
From: Fco. J. Ballesteros <nemo@lsu...>
Subject: 3 questions
Date: Mon, 29 May 2006 19:14:14 +0200

:  After the 'how do we read/store mail on Plan 9' a couple of months ago,
:  has anyone done anything to improve the situation, e.g. memory footprint?

I have done something.
In the Plan B distrib I copied to sources, there's a program that
should work also for Plan 9. It should be at
/n/sources/contrib/nemo/sys/src/cmd/mail2fs/ IIRC

The idea is that mail2fs converts from mbox format to a file tree.
For example, my current inbox (listed from its root), is:

200605/6679/text    rsc@swt...   Re: [9fans] 3 questions                          
200605/6678/text    rsc@swt...   Re: [9fans] 3 questions                          
200605/6677/text    quanstro@quans  Re: [9fans] 3 questions                          
200605/6676/text    jmk@pla...  Re: [9fans] 3 questions                          
200605/6675/text    quanstro@quans  Re: [9fans] 3 questions                          
200605/6674/text    jcenteno@gsyc.  Re: programa en canal historia                   
....

Then, to read mail, I use mails (which merely pretty-prints */[0-9]*/text)
or one for the wrappers. Indeed, cat */[0-9]*/text works well enough.

Would this suffice?
I remember complaints about the speed of using grep to search within this,
but in practice, I manage. If that's a problem, maybe what we need is a
generic file search a-la google for our files.