Re: Long line of OR

> From [email protected]  Thu Oct 13 10:19:06 2011
> Subject: Re: Long line of OR
> From: LuKreme <[email protected]>
> Date: Thu, 13 Oct 2011 09:19:22 -0600
> To: [email protected]
>
> On 12 Oct 2011, at 12:08 , Professional Software Engineering wrote:
> > You may however see more of an improvement by optimizing your regexp:
> > 
> >> (a(a|psumo|dmin)|b(eni|smith)|c(andy|asa|f|pr)|d(b|bl|hosts)|fb|g(lue|ofobo)|hita|ic|l(lc|ogi)|micr|netf|ocb|p(lus|l)|r(smith|oot|cc|k|ocb)|s(cv|nap|pot|u)|t(or|ripa|wit)|v(365|m))
> > 
> > Readability takes a big dive though.
>
> Readability is important, especially since the list changes every now and again.
>
> > Again, youd need to run benchmarks - put recipes in a sandbox, run against a large corpus several times, omitting the longest and shortest times (owing to the effects of cache, etc), and would need to run it on a host which didn't have other demands on it.
>
> OK, I was afraid I was forgetting something like maximum line length or something like that.

There *is* a max line length.  You're nowhere close to it, as the _default_
value is 2048 characters,   And it can be arbitrarily increased by setting
the LINEBUFF variable.   man procmqilrc is your friend

I have recipe conditions that make the one you're 'wondering' about look
like a toy; I've never had to increase LINEBUF, _yet_ anyway. 

Also, for 'readability'` I'm known to do things like:

LIST="( word1    \
      | word2    \
      | word3    \
      | word4    \
      )"
LIST=`echo ${LIST} |tr -d '\t '`

Whereupon, the recipe looks something like:
   :0 
   * ${LIST}
   action


I tend to make -heavy- use of variables in .procmailrc. It greatly
improves the readability of the actual recipies -- you don't have to
reverse-engineer complex regexes _every_time_ you look at the file 
to figure out what theyere doing.

I've got one medium-complex .procmailrc -- for an address that is the
is the  submission address for a moderated newsgroup and a mailing 
list -- 524 lines. It has 101 lines of comments, 75 blank lines, and
108 lines of variable assignements, and 79 recipes with a total of 71
condition tests.  Many of the recipes simply do a variable assignment,
selecting which one of several mailboxes the message will ultimately 
be delivered to.  A few recipes call for more than one action, so there
are chained 'And' recipes with no conditions.