httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Paul J. Reder" <rede...@raleigh.ibm.com>
Subject Re: filtered I/O ordering
Date Thu, 01 Jun 2000 13:22:43 GMT
Greg Stein wrote:
> 1) the hook scheme tries very hard to *avoid* ordering. much of the
>    documentation, the _FIRST and _LAST type stuff, etc is all written with
>    the point of view that ordering is Badness.

Not that I am an expert, but I thought the idea behind the current hook design
was as follows:
   1) Ordering for hooks is usually unimportant so allow easy definition of
      hooks with little or no granularity (NULL predecessor or successor values
      and 3 levels of ordering _FIRST, _MIDDLE, _LAST).
   2) In the event where finer control is required you can specify _MIDDLE+1
      or _MIDDLE-1, etc.
   3) Where ordering is important due to an interdependance between modules
      you can specify predecessors or successors to the hook.

This doesn't seem to avoid ordering, it just makes the assumed lack of
required ordering easy to implement while providing the ability to precisely
control ordering where required.

> 
> 2) I believe that ordering is the typical case and would like to see a
>    model based around that. The hook model is "broadcast" with some extra
>    features to introduce order.

I/O processing may require a different level of ordering than module handlers.
Perhaps part of the confusion in this discussion is from mixing the module
and I/O requirements (I realize this is splitting hairs a little).

-- 

Paul J. Reder

---------------------------------------------------------------------
Noise proves nothing.  Often a hen who has merely laid an egg cackles
as if she laid an asteroid.
		-- Mark Twain
(An apt quote in this political season.)

Mime
View raw message