> Are we better off reading one extra file per request? I guess this
> question may only be answered if we can estimate how often users
> edit their .htaccess files.

"better off reading one extra file per request"?  I wish if it was
true. Currently, all the htacceess files, through all the path to
the file, should be read (of course, only if htaccess is not
disabled). If the file is in the "/htdocs/full/path/to/dir/of/file",
than 7 htaccess files should be read.

As I already wrote, the only complete solution is to avoid (and
disable) the htaccess support. I know that many people, especially
ASPs who host virtual web sites of others in one machine, will not
love this answer, but I only repeat the recommendation of all the
books and documents about tuning Apache.

> My question is whether functionality may be maintained for this
> feature - but at the same time, with less cost. My *guess* is that
> it can, but (as I said) I'm no expert in this field, so I may be
> missing something important.

If you are concerned and want to solve this problem, you are
welcome. But you claim that you are not an expert, so it will be a
little hard. Try to learn Apache internals, and if you have any
solution, you will "buy your world", and many Apache users will
admire you.

In any case, there are already some partial solutions, but I'm not
the man to lecture about them. Again, I joined the thread only to
give the original questioner some ideas about what are the real
bottlenecks of Apache, and to convince him that mod_so is not one
of them. Some of these bottlenecks have solutions, but this was not
the purpose of my original answer. There is a book which focuses on
how to tune Apache. Buy it, and have better answrs for your
questions.

-- 
Eli Marmor

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to