Boris Hollas <[EMAIL PROTECTED]> writes:

> I've made a list of suggestions to make wwwoffle handling easier:

Thank you for taking the time to do this.  It is always interesting to
see what users have to say about WWWOFFLE, even if I don't agree with
everything that is said.  Hopefully my answers will help you
understand why WWWOFFLE is written the way it is even if it is not the
way that you would like it.


> 1. When I edit wwwoffle options and press "update", I subsequently need to 
> press "config" to make changes take effect. It would be easier if "config" is 
> run each time "update" is clicked.

Yes, this might be true.  There are conditions where editing a section
will give an error message.  In the other cases reloading the
configuration automatically might be a good idea.  I think of them as
two separate things which is why they are done this way.

> BTW, what is the "config" button at the end of the page for? Don't I have to 
> run "update" first?

To allow you to update the configuration file.  Perhaps you edited
outside of the WWWOFFLE control page or something.


> 2. When deleting pages from a cache, it would be more convenient if the 
> updated index is shown right after clicking "delete".

But then there would be no chance to see the error messages.


> 3. purge: max-size=0 means that there is no limit. However, I'd expect that 0 
> means that the cache is empty after purging. I'd find the former more 
> intuitive in the form max-size=infinite or -1. Also, a user may want to 
> completely empty the cache, which I do till now by cd <cache>, rm -r *.

Comparing the options max-size and min-free I can see the confusion.
A min-free of 0 means "ignore the limit" which is exactly the same as
saying "leave a minimum of 0 free".  I had not thought about it this
way before.  For min-free therefore using 0 to mean ignore the option
is equivalent to using a value of 0 for the option.

For the max-size option this is not the case.  A max-size of 0 means
"delete all the cached files" (I don't know why anybody would do this)
while I programmed it to mean "ignore the option".

I can change the code to allow max-size=-1 to mean "ignore the
option", but changing max-size=0 to mean "delete all the cached files"
will cause problems for some users.  If people don't change the config
file then it will delete all of their cache.  I just know that this
will cause problems.


> 4. the offline-rerequest problem: I suggest to introduce an option such as 
> "rerequest-cached-pages = yes|no" to replace "pragma-no-cache" and 
> "cache-control-no-cache". Unless there's a need to treat the latter two 
> options separately, it's easier to deal with just one (self-explaining) 
> option to fix rerequesting problems.

Have you tried to fix your browser?

The documentation in the README.CONF file makes it clear when you
should use the pragma-no-cache option.

An option called rerequest-cached-pages is ambiguous because it might
mean that you should re-request all cached pages all the time.  What
you actually want is re-request-cached-pages-when-the-browser-asks or
something.


> 5. The examples for Censor Header are unclear:
> ### Example ###
> # Don't send the username.
> # From = yes
> What does this mean? From=yes makes wwwoffle not send the username?

Yes exactly like the comment and documentation says.

The 'From' header is not used by many browsers, but unless you check
what your browser is sending you might not know that.

> Also, there should be a description for what happens to cookies, e.g. if 
> cookies are enables by wwwoffle but not by the browser.

If cookies are enabled by WWWOFFLE but not by the browser then you
should refer to your browser documentation.  WWWOFFLE has no idea what
the browser might do if it is allowed to see cookies.

Question 3.5 in the FAQ details what might happen if WWWOFFLE caches
pages with cookies when a browser would not.


> 6. If I enter sites to be fetched with http://localhost:8080/refresh-options/ 
> and push "fetch now", wwwoffle doesn't tell when fetching is over (as it does 
> for wwwoffle -fetch).

This is because recursive fetching is performed in the background.  It
might take a long time and behaves differently online and offline.


> 7. wwwoffle should go online before fetching if it is offline.

There are separate options for online and fetching.  It is not
expected that 'wwwoffle -online/-offline/-fetch' will be run manually,
but will be used in automatic scripts.  In this case there is no
problem with running the program twice.


> 8. The wwwoffled.rc script in /contrib/redhat2 is not easy to find and people 
> probably don't know where to put it and what links to create for sys v init. 
> Some pointers in the faq would be helpful.

The FAQ is for WWWOFFLE questions while the files in the contrib
directory are not part of WWWOFFLE, they are just things that people
might find useful.  If the documentation in contrib/README is not
enough then I can remove the wwwoffled.rc script that you have
problems with.

-- 
Andrew.
----------------------------------------------------------------------
Andrew M. Bishop                             [EMAIL PROTECTED]
                                      http://www.gedanken.demon.co.uk/

WWWOFFLE users page:
        http://www.gedanken.demon.co.uk/wwwoffle/version-2.7/user.html

Reply via email to