[ejabberd] ejabberdctl.template patches

Badlop badlop at gmail.com
Tue Jan 5 23:11:59 MSK 2010


2009/12/30 Martin Langhoff <martin.langhoff at gmail.com>:
> -- I think it is worth considering applying these 2 patches in
> ejabberd SVN. See below
> Make concurrent calls to ejabberctl work correctly -- using a bound
> number of atoms:
> http://dev.laptop.org/git/users/martin/ejabberd-xs.git/tree/ejabberdctl_multiconnection.patch?id=5f257b944d71c318bcf25a6da92ca60f4bc42c83

You have summarized the patches and their main drawbacks in
https://support.process-one.net/browse/EJAB-1153

But there is still another metric you forgot when evaluating the
different patches: the new requirements they introduce.

Badlop's patch requires:
  * date
Konstantin Khomoutov's patch requires:
  * nothing new
Martin Langhoff's patch requires:
  * flock
  * runuser
  * bash
  * arithmetic expressions in the shell: are those supported by POSIX?
  * /var/lock/ejabberd creation: who takes care to create it?

Let's assume the patch works correctly when all the requirements are satisfied.
Does the benefit provided by the patch pay the additional requirements?
In your case it does, but does it also for general ejabberd deployments?

What new requirements are introduced in the different operating
systems where ejabberd is commonly used, and are they available in
those operating systems?
(I guess we can't consider Microsoft Windows as an "operating system"
at all for that case).


---
Badlop
ProcessOne


More information about the ejabberd mailing list