This is a weird error. The unison package ( http://www.cis.upenn.edu/~bcpierce/unison/ ) experimental version is 2.12.2, but the unison website latest beta version is 2.12.0 (here: http://www.cis.upenn.edu/~bcpierce/unison/download/beta-test/latest/ ) When I try to communicate from other machine to my gentoo machine, unison complains that they don't have the same version.
Same here: I try to contact the unison server (gentoo, net-misc/unison-2.12.0) with the binary from the site (linux unison-2.12.15-gtk2-text) and I get this: Received unexpected header from the server: expected "Unison 2.12.15\n" but received "Unison 2.12.0\n\000\000\000\000", which differs at "Unison 2.12.0". This can happen because you have different versions of Unison installed on the client and server machines, or because your connection is failing and somebody is printing an error message, or because your remote login shell is printing something itself before starting Unison. (all x86)
Unison needs the same versions on both sides, there's nothing I can do about it. Besides, there is a 2.12.0-linux-gtk2 version here too: http://www.cis.upenn.edu/~bcpierce/unison/download/unison-contributed-binaries/linux/. In the actual CVS tree there are now versions 2.9.1, 2.12.0 and 2.13.0. It used to be a 2.12.2 version which fixed some bugs in 2.12.0, but now we have 2.13.0.
(In reply to comment #2) > Unison needs the same versions on both sides, there's nothing I can do about > it. > Besides, there is a 2.12.0-linux-gtk2 version here too: > http://www.cis.upenn.edu/~bcpierce/unison/download/unison-contributed-binaries/linux/. > In the actual CVS tree there are now versions 2.9.1, 2.12.0 and 2.13.0. It used > to be a 2.12.2 version which fixed some bugs in 2.12.0, but now we have 2.13.0. What I was trying to communicate was that I believe the nulls at the end of the identification header are not supposed to there, I never saw them in the binaries. I could be wrong, but I think as long as it would be "Unison 2.12.0\n" it would work with 2.12.15 too.
> > What I was trying to communicate was that I believe the nulls at the end of the > identification header are not supposed to there, I never saw them in the > binaries. I could be wrong, but I think as long as it would be "Unison 2.12.0\n" > it would work with 2.12.15 too. Nope, it doesn't. Sorry, my fault.