New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
(Deferred) Mininet patches from Big Switch #2
Conversation
when building the package
…ttern in the command output
an unprivileged Python process
like SIGINT (this wasn't a problem before the previous change which ran each subprocess with sudo; for some reason SIGINT behaves itself when the shell is spawned directly); set the shell prompt to ASCII 127 to make it easy to find the end of each command; echo a serial number to ensure the output of multiple commands is synchronized with the caller
spawned in a pseudo-tty
Switch class so tests don't have to guess the dpid
The chunk of dupes at the end are gone, so this is getting closer. Still, These commits aren't in github master or testing, but they are in a branch Bob, what's the status of this code? Thanks, On Tue, Oct 25, 2011 at 3:58 PM, Bob <
|
We have been running our changes internally for some time. Its not clear to On Tue, Oct 25, 2011 at 4:14 PM, Brandon Heller <
|
@brandonheller Regardless, as I said in the original pull message, let's discuss the bigswitch changes here rather than getting derailed by the ovsu changes, unless you think there's a good reason to do so. @mscohen02 |
On Tue, Oct 25, 2011 at 4:25 PM, mscohen02 <
Thanks Mike. It's no big deal either way, this is minor stuff. There's just a chunk of commits from Bob that didn't make it from the I'm trying to find out if this chunk is good code that should get merged now Bob's chunk of commits relates to exposing the built-in OVS controller and thanks,
|
@lantz Thanks Bob, I know I'm asking without doing today - but I'm working to get Ali Y up to speed here, and in his critical path. |
@brandonheller On Oct 25, 2011, at 4:40 PM, Brandon Heller wrote:
|
As long as the code makes it in, in good shape, I'm happy. Cherry-picking works fine when it isn't a ton of commits, like this . Let's do that, like you originally suggested, and separately figure out whether the userovs commits should get merged. I'll review the BS code. As for "large amount of extra work", frequent rebasing is a fact of life with git when you have a bunch of parallel things going on and need to periodically merge them in, and want the history to stay sane. I can do this for ovs if you think it's too much work. |
Actually now that I realize what happened it wasn't terribly difficult to push a feature branch for userovs to github. It is at: On Oct 25, 2011, at 4:53 PM, Brandon Heller wrote:
|
Cool, then do you want to send a pull request to mininet/mininet and cc me? I think we should get in the habit of having someone else review all -b On Tue, Oct 25, 2011 at 6:20 PM, Bob <
|
Not a bad idea; done. On Oct 25, 2011, at 6:34 PM, Brandon Heller wrote:
|
The main other useful thing for this patch is the pty support, which should probably be split out in its own patch and merged. |
I'm removing the 2.0.0 tag as I think this can be deferred to post 2.0.0. |
I believe the resolution of this request is:
|
Fix installing LINC-Switch after modifying rebar configuration
So, here is the first batch of mininet patches from big switch; normally they would create the pull request, but to get things started and to facilitate some discussion and code review, I've created the pull request myself.
This is (as far as I can tell) revised to have the correct rebase, though it seems to include my commits from the userovs branch on yuba. I'd suggest _discussing (only) the big switch changes here, and discussing the userovs changes at: #4._
I haven't reviewed everything, but a couple of things I've seen so far look pretty reasonable, including using sudo so Mininet is only root for the root-like things it needs to do (and so it can be imported into average python code not running as root), and a fix for the annoying control-c-kills-your-xterms (and other subprocesses) bug (assuming that the fix is good and works.)
Although this is a mass of patches, we may want to cherry-pick specific commits.
I'm (hopefully) cc'ing Ed and Mike from big switch so they can join the discussion.
/cc @eswierk @mscohen02
/cc @jvimal @nikhilh