Ask Bjørn Hansen (a...@develooper.com)

Tak Rasmus!

Could you make the diff with "diff -u"?


   - ask

Oh - your performance with qpsmtpd will certainly be better than what you have now then!

> In fact, some of the policy filtering is hidden in the greylisting
> daemon, because there it's perl and doesn't fork.

... and the code cleaner.  :-)

>>> 4. Is this the right place for bug reports about included plugins?
>>
>> Yup! Best form of bug reports are fixes in your github repository though. We also have a little-used bug tracker at http://code.google.com/p/smtpd/issues/list
>
> I've got to understand things a bit more before that ...


Obviously. :-) I'm just pointing it out because for many users qpsmtpd is more of a development platform than an application; so if you are going to do work based on it then you'll be better off starting with a git clone than a tar ball.


  - ask

On Feb 17, 2010, at 3:41, Jost Krieger wrote:

> 1. How many SMTP sessions can be expected on commodity hardware under
>   Linux (RedHat), with prefork? 

If you don't do any heavy processing (spamassassin or virus scanning basically) then ~20/second (or whatever a million day works out to when your server is busiest) shouldn't be a problem.

If you do spamassassin and such, that will be the resource intensive part. qpsmtpd might help over other MTAs though in filtering more of the mail in cheaper ways first.

> 2. How fast is this (I've read Chris Lewis saying 300 sessions/sec)?

That doesn't sound unreasonable, but as soon as you add the spam filtering it goes down -- imagine even just 5 DNS lookups for blacklists and such and now you're also doing 1500 DNS lookups a second.

[...]
> 4. Is this the right place for bug reports about included plugins?

Yup! Best form of bug reports are fixes in your github repository though. We also have a little-used bug tracker at http://code.google.com/p/smtpd/issues/list


  - ask

On Feb 6, 2010, at 12:41, Larry Nedry wrote:

> Here's my patch for require_resolvable_fromhost. As Chris Lewis =
suggested,
> delaying rejection until the RCPT TO hook solved the problem.


Thank you for making a git repository with the change; I pulled it into =
my branch.


- ask=

On Jan 18, 2010, at 8:32, Peter J. Holzer wrote:

> In mid-November I found a bug in Qpsmtpd::config. I created issue 29 on
> Google Code (http://code.google.com/p/smtpd/issues/detail?id=29) and
> some time later checked a fix into my github repository and

Hi Peter!

You did the right thing; except it's better to send patches to the list than pull requests. (So have it in github, but also send the patches to the list for review).

It's basically just:

git format-patch HEAD^.. # or HEAD~5..HEAD~3 or whatever you need to specify the patch series
git send-email 000*

I think my concern with your change was if it was the right thing to do for configuration plugins (as you noted in a comment in your code, too).


- ask

On Dec 28, 2009, at 13:21, Larry Nedry wrote:

> I'm installing qpsmtpd 0.83 on a new server running CentOS 5.4. When =
I run
> make test I get an unexpected message, "command not performed". Is =
this an
> issue that I need to resolve?

Yeah - the test passes so all is well.  I think it's just a test that =
checks that something's not supposed to work (and it doesn't).


- ask=

I think you could make config_dir method in Apache::Qpsmtpd that'll return a directory name if the appropriate PerlSetVar is specified (and fall back to $self->SUPER::config_dir(@_) if not).


- ask=

On Nov 29, 2009, at 16:25, Rick wrote:

Hi Rick!

I'm pretty sure Charlie wasn't just being stubborn but rather just =
trying to figure out what the behavior really should be so we don't =
change it to work compatibility with alpine just to break it with =
something else.

It's entirely possible that the code in question is wrong; I'm guessing =
Charlie was just surprised (as I was) because it's 5 years old and is =
used by many without these problems.


- ask=

Either make a github account, clone my repository and then:

git clone git@github.com:msergeant/qpsmtpd.git
git add new.file another-file.pl
git commit
git push origin

... and then anyone can work with it from your repository. This is how most patches are processed now.

or:

git clone git://github.com/abh/qpsmtpd.git
git add new.file another-file.pl
git commit
git format-patch HEAD^..
git send-email --to=qpsmtpd@perl.org 0001*

to manage the changes in git and use git to send them to the list.

... or of course you could just send them to the list or to me or put them on http somewhere and send a link -- but since you asked. :-)


- ask

I'm not exactly sure what your question is, but

http://search.cpan.org/search?query=anyevent

shows a sample of available event implementations and other modules.


- ask

On Nov 23, 2009, at 11:41, Matt Sergeant wrote:

> Is anyone interested in an AnyEvent mode Qpsmtpd? I have the code written (mostly hacked right now, but should work).
>
> In theory it might be faster than the Danga::Socket based one, and AnyEvent seems to receive regular updates more than Danga::Socket these days.

There are also more "other stuff" available with AnyEvent -- I think it'd be cool!


- ask=

On Nov 5, 2009, at 10:31, Shad L. Lords wrote:

> This might not be the best place to fix this. The correct way would
> be to fix header->add to always stick the headers at the top.

Sounds good to me.  Would that mess up anything else?

While Robert is right that this is much easier to work with as a  
proper patch; this one was easy enough that after staring at it for  
long enough to see the typo I applied it; thanks!


  - ask

Try using the 'swaks' tool to test email instead: http://jetmore.org/john/code/#swaks
http://jetmore.org/john/code/swaks

(Although we should fail more gracefully when you give it an invalid  
email message...)


  - ask

On Oct 16, 2009, at 14:25, Steve Kemp wrote:

> Maybe ask doesn't get much SPAM thanks to qpsmtpd, but
> it might make sense ask people to send mail to this list:

Yeah; the link is more helpful ([email protected: a...@develooper.com] gets an insane
amount of spam; it surprises me that email is still functional at all).


  - ask

On Jul 20, 2009, at 9:45, Stefan Priebe wrote:

[timeouts not working]
> oh i'm really sorry - it was caused by a module config_cache which i
> found somewhere in the net. When i disable it it works fine :-(

Weird -- what does that module/plugin do?  It shouldn't make the  
timeouts not work ...  (the timeout default to 20 minutes if it's not  
specified otherwise).


  - ask

Weird - the code (lib/Qpsmtpd/TcpServer/Prefork.pm#read_input) is  
pretty straightforward.  Maybe some goofy safe signals issue?

Most of the code in that method is shared with lib/Qpsmtpd/
TcpServer.pm -- patches to unify them better would be nice, too.


  - ask

On Jul 15, 2009, at 2:18, Robin Bowes wrote:

[pushing to github]
> I already did! :)

Thanks.  I'll look at it next week.


  - ask

Just clone the repository on github and then go crazy in your "fork".   
When you are done we'll figure out to integrate the changes into "my"  
repository.


  - ask

+1 on getting these things into git.   (Likewise for the debian stuff  
if that makes sense).


  - ask