net-analyzer/mtr-0.85 upgraded from 0.82 built with USE="-gtk -ipv6 -suid" fails to run and returns to a prompt, but the terminal is unusable. mingdao@router ~ $ eix net-analyzer/mtr [I] net-analyzer/mtr Available versions: 0.82 0.85 {gtk ipv6 suid} Installed versions: 0.85(10:54:58 AM 08/30/2013)(-gtk -ipv6 -suid) Homepage: http://www.bitwizard.nl/mtr/ Description: My TraceRoute, an Excellent network diagnostic tool mingdao@router ~ $ zgrep ipv6 /proc/config.gz mingdao@router ~ $ zgrep -i ipv6 /proc/config.gz # CONFIG_IPV6 is not set mingdao@router ~ $ su - Password: router ~ # mtr www.google.com (After issuing the above command the terminal must be reset. See URL of screenshot. I can't seem to attach the screenshot in this bug report.) Reproducible: Always Steps to Reproduce: 1.router ~ # mtr www.google.com Actual Results: router ~ # eix net-analyzer/mtr [I] net-analyzer/mtr Available versions: 0.82 0.85 {gtk ipv6 suid} Installed versions: 0.85(10:54:58 AM 08/30/2013)(-gtk -ipv6 -suid) Homepage: http://www.bitwizard.nl/mtr/ Description: My TraceRoute, an Excellent network diagnostic tool router ~ # mtr www.google.com The terminal output is messed up (see attached screenshot) and the cursor in the vt does not move when text is input. If "reset" is issued, it will reset the term. Expected Results: Expected a graph of traceroute as usual. Might be related to https://github.com/traviscross/mtr/issues/19 but I have not tried to patch it to see.
Created attachment 357926 [details] screenshot of mtr failure This is a screenshot of the mtr-0.85 failure; mtr-0.82 still works on all the effected machines.
Reproducible here.
What if you try with the --curses switch?
(In reply to Jeroen Roovers from comment #3) > What if you try with the --curses switch? mtr --curses www.google.com produces the same result. It's built -gtk so I assumed the interface is ncurses automatically, but tried at your suggestion.
May by reasonable introduce 9999 ebuild with latest git? Or bump patch for this bug from from https://github.com/traviscross/mtr/commit/12c53f98e44598b87d3f2308e0d892f49d7af8e4
(In reply to Vladimir Varlamov from comment #5) > May by reasonable introduce 9999 ebuild with latest git? Or bump patch for > this bug from from > https://github.com/traviscross/mtr/commit/ > 12c53f98e44598b87d3f2308e0d892f49d7af8e4 Interesting, but that's not a proper solution to getting the current stable version working. The solution is to set USE=ipv6.
(In reply to Jeroen Roovers from comment #6) > (In reply to Vladimir Varlamov from comment #5) > > May by reasonable introduce 9999 ebuild with latest git? Or bump patch for > > this bug from from > > https://github.com/traviscross/mtr/commit/ > > 12c53f98e44598b87d3f2308e0d892f49d7af8e4 > > Interesting, but that's not a proper solution to getting the current stable > version working. The solution is to set USE=ipv6. IPv6 isn't available many places, and having mtr-0.82 is far less obnoxious at this point that any USE=ipv6 or CONFIG_IPV6=y.
*** Bug 486728 has been marked as a duplicate of this bug. ***
Are we waiting on upstream? Is anyone testing upstream commits? I just assumed so, but if help is needed, I can test this coming week.
I tested net-analyzer/mtr-9999 (thanks for ebuild) as My traceroute [v51dafcae9013c82021ffb91d851f80486e42b89f] "Unable to allocate IPv6 socket..." still here but the program as a whole works I think it necessary to wait for the next 0.86 release
v0.85+git:99dbe3f1 works with USE="-gtk -ipv6 -suid" Could we get a 0.86 pkg in ~arch to test?
(In reply to Bruce Hill from comment #11) > v0.85+git:99dbe3f1 works with USE="-gtk -ipv6 -suid" > > Could we get a 0.86 pkg in ~arch to test? Alternatively, you could ask upstream for a release.
(In reply to Jeroen Roovers from comment #12) > (In reply to Bruce Hill from comment #11) > > v0.85+git:99dbe3f1 works with USE="-gtk -ipv6 -suid" > > > > Could we get a 0.86 pkg in ~arch to test? > > Alternatively, you could ask upstream for a release. Gosh, so sorry for the static. Recent use of tmux, then i3, on a laptop has me watching tiny windows/panes/workspaces and I thought there was v0.86 on Travis' GitHub. Will just keep using the git version until the release.
Should be fixed in 0.86 which has landed. Please test.
at x86 and amd64 works well
(In reply to Jeroen Roovers from comment #14) > Should be fixed in 0.86 which has landed. Please test. Fixed. You can close this bug.
Thanks.
