<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Jun 8, 2016 at 7:55 PM, john hood <span dir="ltr"><<a href="mailto:cgull@glup.org" target="_blank">cgull@glup.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">Yeah, would you open a GitHub issue?</blockquote><div><br></div><div>Will do. </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<br>
I tried my two Macs, and it breaks for me, but differently: they get<br>
link-local IPv6 addresses from mDNS, requiring an interface identifier,<br>
so mosh-client doesn't know where to send packets.<br>
<br>
I'm wondering how you're getting global IPv6 addresses from mDNS, even<br>
though both your client and server are on the same local network.<br>
Judging by the hostnames you're using and the IPv6 address, this is a<br>
corporate OS X install on $yourverylargecompany's network.<br>
<br></blockquote><div>Yes, good guess. :-) I think it's relatively common to have a global v6 address allocated to a system which is RFC1918 in v4 these days.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
If the server involved has regular DNS A/AAAA records, does that work<br>
with Mosh?<br></blockquote><div><br></div><div>I verified I see the same behavior moshing into systems with both global A/AAAA records, running an older version of mosh server (mosh 1.2.4a). So it seems it isn't dependent on mDNS, just an old server running on a dual stack server. I'll write up the bug.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<br>
regards,<br>
<br>
--jh<br>
<span class=""><br>
On 6/8/16 9:37 PM, John Hood wrote:<br>
> Er, that's an interesting bug, it should work for either v4 or v6. I'll<br>
> check mDNS tonight.<br>
><br>
> regards,<br>
><br>
> --jh<br>
><br>
> On 06/08/2016 19:54, Bryce Bockman wrote:<br>
>> I installed the OS X binary version (gotten from<br>
</span>>> <<a href="https://github.com/mobile-shell/mosh/releases/download/mosh-1.2.5.95rc1/mosh-1.2.5.95rc1.pkg" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/mosh-1.2.5.95rc1/mosh-1.2.5.95rc1.pkg</a>><a href="https://github.com/mobile-shell/mosh/releases/download/mosh-1.2.5.95rc1/mosh-1.2.5.95rc1.pkg" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/mosh-1.2.5.95rc1/mosh-1.2.5.95rc1.pkg</a>).<br>
<span class="">>> The install went fine, and I was able to mosh into an older server<br>
>> running on stock Ubuntu 14.04.<br>
>><br>
>> Found an interesting bug/feature, likely due to ipv6 related changes.<br>
>> I often use mDNS name resolution. When moshing from the release<br>
>> candidate to the older server mentioned above using mDNS target I get<br>
>> the following:<br>
>><br>
>> bb@bb-mac:~$ mosh beb-dev01.local<br>
>><br>
>> mosh-server: Bad IP address (2606:b400:605:d0a0:dacb:8aff:fe69:7a71)<br>
>><br>
>> Usage: mosh-server new [-s] [-v] [-i LOCALADDR] [-p PORT[:PORT2]] [-c<br>
>> COLORS] [-l NAME=VALUE] [-- COMMAND...]<br>
>><br>
>> Connection to beb-dev01.local closed.<br>
>><br>
>> /usr/local/bin/mosh: Did not find mosh server startup message.<br>
>><br>
>> bb@bb-mac:~$<br>
>><br>
>><br>
>> Note that mDNS resolves either v4 or v6 addresses:<br>
>><br>
>><br>
>> bbockman@bbockman-mac:~$ ping6 beb-dev01.local<br>
>><br>
>> PING6(56=40+8+8 bytes) 2606:b400:605:d0a0:16b1:c8ff:fe00:5da0 --><br>
>> 2606:b400:605:d0a0:dacb:8aff:fe69:7a71<br>
>><br>
>> 16 bytes from 2606:b400:605:d0a0:dacb:8aff:fe69:7a71, icmp_seq=0<br>
>> hlim=64 time=0.549 ms<br>
>><br>
>> bbockman@bbockman-mac:~$ ping beb-dev01.local<br>
>><br>
>> PING beb-dev01.local (10.145.137.58): 56 data bytes<br>
>><br>
</span>>> 64 bytes from 10.145.137.58 <<a href="http://10.145.137.58" rel="noreferrer" target="_blank">http://10.145.137.58</a>>: icmp_seq=0 ttl=64<br>
<span class="">>> time=0.359 ms<br>
>><br>
>><br>
>> If I explicitely mosh using the v4 address things work (mosh<br>
>> 10.145.137.58).<br>
>><br>
>> Not sure the extent of the behavior. e.g. Would the same behavior be<br>
>> seen with an older server that has a AAAA record with newer client or<br>
>> not. Should open an issue on github?<br>
>><br>
>><br>
>><br>
>> On Wed, Jun 8, 2016 at 12:44 PM, Keith Winstein <<a href="mailto:keithw@mit.edu">keithw@mit.edu</a><br>
</span><span class="">>> <mailto:<a href="mailto:keithw@mit.edu">keithw@mit.edu</a>>> wrote:<br>
>><br>
>> Thanks all who have tested the mosh 1.2.6 release candidate so far.<br>
>><br>
>> We're still looking for a postive ACK that somebody has<br>
>> successfully tested the release candidate on:<br>
>><br>
>> - Fedora<br>
>> - OS X binary .pkg<br>
>> (<a href="https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg</a>)<br>
>> - OS X compiling from source<br>
>><br>
>> Thanks all,<br>
>> Keith<br>
>><br>
>> On Sun, Jun 5, 2016 at 8:39 PM, Richard Woodbury<br>
</span><span class="">>> <<mailto:<a href="mailto:rpwoodbu@mybox.org">rpwoodbu@mybox.org</a>><a href="mailto:rpwoodbu@mybox.org">rpwoodbu@mybox.org</a>> wrote:<br>
>><br>
>> I have confirmed that the RC builds and runs OK in Mosh for<br>
>> Chrome.<br>
>><br>
>> On Thu, Jun 2, 2016 at 11:47 PM Richard Woodbury<br>
</span><span class="">>> <<mailto:<a href="mailto:rpwoodbu@mybox.org">rpwoodbu@mybox.org</a>><a href="mailto:rpwoodbu@mybox.org">rpwoodbu@mybox.org</a>> wrote:<br>
>><br>
>> Hi, Keith. I don't have predictable time right now, but as<br>
>> I can, I'll look into this for Mosh for Chrome. I did<br>
>> manage to do a quick "smoke test" build, but I'm getting a<br>
>> bunch of protobuf link errors. I'll need to do more<br>
>> investigation, which may include bringing in a newer NaCl<br>
>> SDK so I can also get newer libraries from naclports.<br>
>> Hopefully I'll find time in the next day or two.<br>
>><br>
>> I can add that it builds OK on the Raspberry Pi (ARM,<br>
>> Raspbian), and the performance improvement makes a<br>
</span>>> /huge/ difference! mosh-server used to take most of the<br>
<span class="">>> CPU for busy output or large window size and was notably<br>
>> sluggish, depending on local echo to be usable. Now the<br>
>> CPU usage is in the noise, and it responds wonderfully.<br>
>><br>
>><br>
>> On Mon, May 30, 2016 at 3:34 AM Keith Winstein<br>
</span><span class="">>> <<mailto:<a href="mailto:keithw@mit.edu">keithw@mit.edu</a>><a href="mailto:keithw@mit.edu">keithw@mit.edu</a>> wrote:<br>
>><br>
>> Hi folks,<br>
>><br>
>> Could you please send in some positive testing reports<br>
>> on the Mosh 1.2.6 release candidate?<br>
>><br>
>> It would be great to have an independent "looks good"<br>
>> from at least the following platforms before we cut<br>
>> the release:<br>
>><br>
>> - Fedora<br>
>> - the OS X binary .pkg<br>
>> - OS X compiling from source<br>
>> - OpenBSD<br>
>> - FreeBSD<br>
>> - NetBSD<br>
>> - Chrome<br>
>><br>
>> Thanks,<br>
>> Keith<br>
>><br>
>> On Wed, May 25, 2016 at 8:56 PM, john hood<br>
</span><div><div class="h5">>> <<mailto:<a href="mailto:cgull@glup.org">cgull@glup.org</a>><a href="mailto:cgull@glup.org">cgull@glup.org</a>> wrote:<br>
>><br>
>> Hi all,<br>
>><br>
>> We're happy to announce the upcoming release of<br>
>> Mosh 1.2.6, and are<br>
>> calling for testing on Mosh 1.2.5.95rc1. The<br>
>> release has picked up<br>
>> some minor new features in the year since the last<br>
>> release such as<br>
>> better IPv6 support and tools to handle orphaned<br>
>> sessions. However,<br>
>> it's also seen significant improvements in<br>
>> performance, testing, and<br>
>> portability.<br>
>><br>
>> The Changelog for this release:<br>
>><br>
>> * New features:<br>
>> * Add Travis CI builds for Linux and Mac.<br>
>> (Anders Kaseorg, others)<br>
>> * Add a --local option to run without ssh.<br>
>> (John Hood)<br>
>> * Mosh now returns exitstatus reflecting<br>
>> connection success.<br>
>> (John Hood)<br>
>> * Add a end-to-end test suite and many tests.<br>
>> (John Hood)<br>
>> * Implement timeouts and signals to help<br>
>> address orphaned sessions.<br>
>> (John Hood)<br>
>> * Major rework of Mosh's display<br>
>> differencing/rendering<br>
>> code with much improved performance for slow<br>
>> machines. (John Hood)<br>
>> * Implement ANSI back/forward tab (CSI CBT,<br>
>> CSI CHT).<br>
>> (John Hood)<br>
>> * Do not start user shell until network<br>
>> session starts.<br>
>> (John Hood)<br>
>> * Add options for more flexible specification<br>
>> of IPv4/IPv6<br>
>> hostname resolution. (John Hood)<br>
>> * Improved bash completion. (Steve Dignam,<br>
>> HIGUCHI Yuta)<br>
>> * Add options for different methods of<br>
>> resolving the remote host<br>
>> address, allowing operation without<br>
>> SshProxyCommand. (John Hood)<br>
>><br>
>> * Platform support:<br>
>> * Add configurable support for Apple Common<br>
>> Crypto and<br>
>> Nettle, in place of OpenSSL. Implement<br>
>> base64 locally.<br>
>> (John Hood)<br>
>> * Workaround Cygwin select() bug. (John Hood)<br>
>> * Updates to Debian packaging. (Anders<br>
>> Kaseorg, Keith Winstein)<br>
>> * Workaround a glibc-2.22 issue causing<br>
>> segfaults on Debian Sid.<br>
>> (John Hood with help from many others)<br>
>> * Prefer c++ to g++, for systems like FreeBSD<br>
>> where g++ is not usable.<br>
>> (John Hood)<br>
>> * Fixes for Illumos Hipster 20151003. (John Hood)<br>
>> * Disable -Werror for protobuf code, to<br>
>> resolve a new gcc6 warning.<br>
>> (John Hood)<br>
>> * Link test for -fstack-protector-all on an<br>
>> embedded platform.<br>
>> (Baruch Siach)<br>
>> * Resolve issue with bswap64() on<br>
>> FreeBSD-CURRENT with libc++-3.8.0.<br>
>> (John Hood)<br>
>> * Fix issue with RECVTOS error message on<br>
>> client on FreeBSD.<br>
>> (John Hood)<br>
>><br>
>> * Bug fixes:<br>
>> * Remove an assertion causing aborts on<br>
>> Unicode fallback found by<br>
>> fuzzing with afl. (Keith Winstein)<br>
>> * Fix a server hang with XON/XOFF on BSD<br>
>> systems. (John Hood)<br>
>> * Fix a typeahead-prediction bug that caused<br>
>> display corruption on<br>
>> urxvt. (John Hood)<br>
>><br>
>> Source code is available as<br>
</div></div>>> <<<a href="https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.tar.gz" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.tar.gz</a>><a href="https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.tar.gz" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.tar.gz</a>>.<br>
<span class="">>> The SHA256 sum for this file is<br>
>> a2697c41cfc8c92dc7a743dd101849a7a508c6986b24d6f44711d8533d18fcf5<br>
>><br>
>> One standalone OS X package is available:<br>
>><br>
>> *<br>
</span>>> <<<a href="https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg</a>><a href="https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg" rel="noreferrer" target="_blank">https://github.com/mobile-shell/mosh/releases/download/untagged-0d43a0f73825e9952677/mosh-1.2.5.95rc1.pkg</a>><br>
<div><div class="h5">>> is an i386/x86_64 build for OS X 10.9 and higher.<br>
>> The SHA256 sum for<br>
>> this file is<br>
>> 48a56d83d0ef655d38e0ea596fd9cac98c0dc433cb5356205d26748350d47e6c<br>
>><br>
>> (If you are using a package system such as<br>
>> MacPorts or Homebrew, I<br>
>> recommend using that, though.)<br>
>><br>
>> As always, Ubuntu PPA builds of the latest source<br>
>> are available at<br>
>> ppa:keithw/mosh.<br>
>><br>
>> Packagers, please note that Mosh has some minor<br>
>> dependency changes: Perl<br>
>> is now required to be >= 5.14, but IO::Socket<br>
>> modules are no longer<br>
>> required. If anybody needs to package for older<br>
>> versions of Perl, talk<br>
>> to me and I'll probably bring something into the<br>
>> release. Also, if<br>
>> anyone needs an OS X package for 10.8 or lower,<br>
>> please contact me.<br>
>><br>
>> Your testing is very unlikely to prove Mosh to be<br>
>> free of bugs, but your<br>
>> testing will help us make 1.2.6 a better release.<br>
>> Please report any<br>
>> issues you find on Github, and we can be found on<br>
>> IRC at<br>
>> <irc://<a href="http://irc.freenode.net/mosh" rel="noreferrer" target="_blank">irc.freenode.net/mosh</a><br>
</div></div>>> <<a href="http://irc.freenode.net/mosh" rel="noreferrer" target="_blank">http://irc.freenode.net/mosh</a>>>.<br>
<span class="">>><br>
>> Looking ahead, we expect Mosh 1.3 to be a feature<br>
>> release, bringing<br>
>> significant new functionality. SSH agent<br>
>> forwarding is high on the list.<br>
>><br>
>> The Mosh team thanks you for your help.<br>
>><br>
>> --John Hood<br>
>><br>
>><br>
>> _______________________________________________<br>
>><br>
>> mosh-devel mailing list<br>
</span>>> <a href="mailto:mosh-devel@mit.edu">mosh-devel@mit.edu</a> <mailto:<a href="mailto:mosh-devel@mit.edu">mosh-devel@mit.edu</a>><br>
<span class="">>> <a href="http://mailman.mit.edu/mailman/listinfo/mosh-devel" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/mosh-devel</a><br>
>><br>
>><br>
>><br>
>><br>
>> _______________________________________________<br>
>> mosh-devel mailing list<br>
</span>>> <a href="mailto:mosh-devel@mit.edu">mosh-devel@mit.edu</a> <mailto:<a href="mailto:mosh-devel@mit.edu">mosh-devel@mit.edu</a>><br>
<div class=""><div class="h5">>> <a href="http://mailman.mit.edu/mailman/listinfo/mosh-devel" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/mosh-devel</a><br>
>><br>
>><br>
>><br>
>><br>
>> _______________________________________________<br>
>> mosh-devel mailing list<br>
>> <a href="mailto:mosh-devel@mit.edu">mosh-devel@mit.edu</a><br>
>> <a href="http://mailman.mit.edu/mailman/listinfo/mosh-devel" rel="noreferrer" target="_blank">http://mailman.mit.edu/mailman/listinfo/mosh-devel</a><br>
><br>
<br>
</div></div></blockquote></div><br></div></div>