Re: ircd+enge5 patch released

[grega@xxxxxxxxxx]

Hi.

First of all: critics should be constructive...

I think that the whole idea of a bunch patch is pretty cool.

Some things may be a break of users privacy and shouldn't be run 
(at least not on ircnet.)
But on some other networks (that use the same irc daemon) the 
options might be usefull (depends on the "politics" there)
However the included qpi patch has settings that default for a network
like ircnet. I wouldn't include that patch in enge5.

[fiction@kronik irc2.10.3p1]$ patch -p1 < irc2.10.3p1-enge5.patch 
patching file common/common_def.h
patching file common/match.c
patching file common/msg_def.h
patching file common/numeric_def.h
patching file common/os.h
patching file common/parse.c
patching file common/send.c
patching file common/struct_def.h
patch: **** malformed patch at line 388:  typedef       struct  Channel
aChannel;

I tried on a completely clean 2.10.3p1 ircd and it doesn't apply.
Some lines in the patch are broken. - let's try this one:
http://below.ijs.si/~grega/irc/irc2.10.3p1-enge5.patch.new.gz
MD5 (irc2.10.3p1-enge5.patch.new.gz) = 2771bd846b11158eed2f7a2d9ce5af4d

I also had problems - the stuff requies openssl (for md5) and does no check
for it.
I would modify configure to search for it or something - not just 
#include <openssl/md5.h>.

Then the USERHOST_SCRAMBLE.

It can't be checked via stats D if this option is enabled or not.

fiction =f528764d.915e786e.0.1 * :neki
looks quite good...
but then you can do stats L and the whole md5 stuff becomes useless
(notice that however by default stats L is disabled for non-opers)
fiction[fiction@xxxxxxxxx] 0 23 1 8 0 :52

915e786e is part of the md5 hash of the servername - why is it good to 
include such a thing?

I would prefer some user flag - like +x to turn that option on and off. 
Bans (and +e and +I) should work for both - the scrambeled host and for the
real.
And lets say a local oper can always see the real host if he is doing
whois, 
userhost or something.

c'ya,
fiction

This is a bit offtopic: 
...
>Note that this is the first public release of this patch and its already
>used by a wide range of major servers on IRC.

The whole thing sounds like a tv commercial. :-)
The features of a product should be the reason to try it not just the fact
that so 
much people are already using it.