Message ID | 1496082575-24450-1-git-send-email-glbsalazar@gmail.com |
---|---|
State | Changes Requested |
Headers | show |
Hello, On Mon, 29 May 2017 19:29:35 +0100, Gonçalo Salazar wrote: > Added notes to the mosh package help to clarify some > issues related with it to ensure mosh will work properly after adding it. > This includes adding a proper LOCALE on the host and an extra flag when using > mosh with dropbear. The commit title should start with the name of the package, i.e: mosh: add notes to clarify runtime issues > https://mosh.mit.edu The notes should go before the upstream URL, and the indentation should be one tab + two spaces. > + > + Mosh requires a host LOCALE matching the client provided LOCALE. I'm sorry, but I don't understand this sentence. > + When using mosh with dropbear, a version with the --no-ssh-pty option on > + the client is required. And also this one. So for me, those explanations do not clarify anything :-/ Best regards, Thomas
Hello, Thank you for the inputs, I realize that the notes are not really clear. I'm trying to convey the idea of runtime dependencies that exist if you are trying to include mosh in a host you will be accessing later. I'll write the notes. Should I resubmit the patch after clarifying the notes? Best regards, Gonçalo 2017-05-29 21:39 GMT+01:00 Thomas Petazzoni <thomas.petazzoni@free-electrons.com>: > Hello, > > On Mon, 29 May 2017 19:29:35 +0100, Gonçalo Salazar wrote: >> Added notes to the mosh package help to clarify some >> issues related with it to ensure mosh will work properly after adding it. >> This includes adding a proper LOCALE on the host and an extra flag when using >> mosh with dropbear. > > The commit title should start with the name of the package, i.e: > > mosh: add notes to clarify runtime issues > >> https://mosh.mit.edu > > The notes should go before the upstream URL, and the indentation should > be one tab + two spaces. > >> + >> + Mosh requires a host LOCALE matching the client provided LOCALE. > > I'm sorry, but I don't understand this sentence. > >> + When using mosh with dropbear, a version with the --no-ssh-pty option on >> + the client is required. > > And also this one. > > So for me, those explanations do not clarify anything :-/ > > Best regards, > > Thomas > -- > Thomas Petazzoni, CTO, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Hello, On Mon, 29 May 2017 22:02:08 +0100, Gonçalo Salazar wrote: > Thank you for the inputs, I realize that the notes are not really clear. > I'm trying to convey the idea of runtime dependencies that exist if > you are trying to include mosh in a host you will be accessing later. Beware that in the context of Buildroot, "host" is often used to indicate the machine where the Buildroot build is done, as opposed to the "target", where the stuff built by Buildroot is executed. Hence, it should be made extra clear that you mean "host" in the networking sense, and not the build sense. > Should I resubmit the patch after clarifying the notes? Yes, please. Thanks! Thomas
Hello, Just to ensure that the patch is submitted correctly this time, is this clearer? Mosh requires that the remote machine has a locale setting matching the one on the client. When using mosh with dropbear on a remote machine, the client needs mosh with a version above 1.3.0 and the --no-ssh-pty option needs to be passed when starting the connection. Thanks, Gonçalo Salazar 2017-05-29 22:15 GMT+01:00 Thomas Petazzoni <thomas.petazzoni@free-electrons.com>: > Hello, > > On Mon, 29 May 2017 22:02:08 +0100, Gonçalo Salazar wrote: > >> Thank you for the inputs, I realize that the notes are not really clear. >> I'm trying to convey the idea of runtime dependencies that exist if >> you are trying to include mosh in a host you will be accessing later. > > Beware that in the context of Buildroot, "host" is often used to > indicate the machine where the Buildroot build is done, as opposed to > the "target", where the stuff built by Buildroot is executed. > > Hence, it should be made extra clear that you mean "host" in the > networking sense, and not the build sense. > >> Should I resubmit the patch after clarifying the notes? > > Yes, please. > > Thanks! > > Thomas > -- > Thomas Petazzoni, CTO, Free Electrons > Embedded Linux and Kernel engineering > http://free-electrons.com
Hello, On Mon, 29 May 2017 22:30:29 +0100, Gonçalo Salazar wrote: > Just to ensure that the patch is submitted correctly this time, is this clearer? > > Mosh requires that the remote machine has a locale setting matching > the one on the client. > > When using mosh with dropbear on a remote machine, the client needs mosh > with a version above 1.3.0 and the --no-ssh-pty option needs to be passed > when starting the connection. Yes, much better! Thanks! Thomas
diff --git a/package/mosh/Config.in b/package/mosh/Config.in index e0ff2cf..9f09b4f 100644 --- a/package/mosh/Config.in +++ b/package/mosh/Config.in @@ -25,3 +25,7 @@ config BR2_PACKAGE_MOSH of user keystrokes. https://mosh.mit.edu + + Mosh requires a host LOCALE matching the client provided LOCALE. + When using mosh with dropbear, a version with the --no-ssh-pty option on + the client is required.
Added notes to the mosh package help to clarify some issues related with it to ensure mosh will work properly after adding it. This includes adding a proper LOCALE on the host and an extra flag when using mosh with dropbear. Signed-off-by: Gonçalo Salazar <glbsalazar@gmail.com> --- package/mosh/Config.in | 4 ++++ 1 file changed, 4 insertions(+)