Asterisk supports linkage against Corosync, but that seems deliberately
avoided in the Debian packaging.

I wonder why it was avoided - changelog entry for 1:13.0.0~dfsg-3 says
"Remove corosync support: doesn't work" but doesn't really explain what
didn't work.

If it was bug#771432 then that bug seems to only suggest disabling the
module by default, not avoiding building the module altogether (might
work fine when properly configured).  Also, if it is that issue, then
that bugreport should probably be closed, referencing that release.

Perhaps the seemingly major cleanup changes to Corosync 2.3.5-2 solves
the reviously experienced issues?

 - Jonas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=KKwf
-----END PGP SIGNATURE-----

"Doesn't work" is #771432 which I didn't have time to properly check.
But also the fact that it is now disabled by default in the upstream
configuration.

Do you use it? If there's a need for it, I'll try to take the time to
research it.

Quoting Tzafrir Cohen (2015-11-08 05:10:55)

I don't use it myself currently, and don't ask to do research about it.
Currently¹ the feature will be enabled or not based on the existance of
the libcpg-dev in the development environment.  That's unstable.

Your reasoning shared here sounds sensible to me.

I suggest to a) document the reasoning as a comment in debian/rules next
to b) explicitly disabling the feature (I guess --without-corosync would
work but am not sure) to avoid uncertainty both in code and in following
intend of code.


 - Jonas

¹ In my latest test builds, at least - you now mention that defaults
have changed, but I still suggest to disable explicitly.