Re: [BFilter-users] Signal handler request

[Joseph A. <jos...@ma...>]

On Sunday 31 August 2003 20:48, gypsy wrote:
> I'd like to be able to
>
> killall -USR1 bfilter
>
> and have it reload /etc/bfilter/config.  Better yet, changes made to
> ~config should be dynamic such that each request would check ~config for
> changes before getting the requested document.  That's because I run
> bfilter as a tsr listening on my internal interface 192.168.1.254:8080
> and I may want to proxy that through privoxy (which listens on :8118)
> rather than passing the request on to Apache directly.
Well, the gui version backends actually do this check before each request.
The problem with the standalone version is that backends don't communicate 
with the master process in any way, nor do they read config themselves.
It should not be a problem to reread the config in the master process, but 
then the current backends won't catch the changes until the client connection 
is closed.
BTW, the USR1 signal is already handled by backends.  It tells them to quit as 
soon as the current request and response finish.

> I am confused by "next-hop".  What is its purpose/what does it do?
> Could we please have comments in ~config and/or docs on sourceforge?
Hop is basically a way from one proxy (or a client or a sever) to another.
Probably I should use more simple words, but if I just write "http proxy", 
then the user could be confused: "I thought bfilter IS the http proxy".
If your config doesn't have comments, thats because Windows versions before 
0.8.1 were stripping them when you used the basic configuration dialog.
I don't put the documentation on the web because I feel it needs improving, 
but I am too lazy to do that.

> BFilter is AWESOMELY fast!  It is a bit of a memory pig, but I can live
It doesn't use that much memory.  You can't trust what the system tells you, 
because most of the memory is shared between backends.

> with that.  I had to compile it on a pre 9.0 version of Slackware
> because it won't compile under gcc 3.2.2 (but then you warned me it
> probably wouldn't); it would be nicer if something could be done to get
> gcc 3.x to work.
Compiles fine on gcc 3.2.3 here.  I think it was a problem with a specific 
version of bfilter.

> Q:  Does rules.local override or supplement rules?  IOW, if rules.local
> is empty, does rules still apply?
It supplements them.  Here is what it does exactly:
1. Read a rule block from rules.local.
2. Fill the absent parameters with defaults from rules.local (normally there 
are no defaults there, unless you want to override global defaults).
3. If rules contains a block with the same pattern, use this block to fill the 
parameters which are still absent.
4. Fill the absent parameters with defaults from rules.
5. Fill the absent parameters with the hardcoded defaults.

> If there is a bash/perl/Etc. guru out there who wants to write a
> conversion program that reads privoxy's *.action files and writes
> rules.local, that would be A Very Good Thing for a ~/contrib directory
> (hint hint).
Is there a need in that?  My rules.local is currently empty and I see ads so 
rarely that I don't bother to create rules for them.