Bugzilla – Full Text Bug Listing
|
Description
Andreas Sewe
(In reply to Andreas Sewe from comment #0) > If I understand David's comment [2] in IPZilla correctly, we should be using > 0.1.53 for Mars.1, right? Cc'ing David to hear his take on this. (In reply to Andreas Sewe from comment #1) > (In reply to Andreas Sewe from comment #0) > > If I understand David's comment [2] in IPZilla correctly, we should be using > > 0.1.53 for Mars.1, right? > > Cc'ing David to hear his take on this. Yes, definitely should use 1.53 for Mars.1 (we in the Platform do). It is part of the latest "R-build" from Orbit: http://download.eclipse.org/tools/orbit/downloads/drops/R20150821153341/repository FYI, the current EGit contribution (4.0.2) [1] for RC4 apparently still contributes JSch 0.1.51 [2] @EGit team: Is this intentional? [1] <https://git.eclipse.org/c/simrel/org.eclipse.simrel.build.git/tree/egit.b3aggrcon?h=Mars_maintenance&id=b309f64b0f90cef6495f8769f3681907b305dcac> [2] <http://download.eclipse.org/egit/updates-4.0.2/plugins/> (In reply to Andreas Sewe from comment #3) > FYI, the current EGit contribution (4.0.2) [1] for RC4 apparently still > contributes JSch 0.1.51 [2] > > @EGit team: Is this intentional? > > [1] > <https://git.eclipse.org/c/simrel/org.eclipse.simrel.build.git/tree/egit. > b3aggrcon?h=Mars_maintenance&id=b309f64b0f90cef6495f8769f3681907b305dcac> > [2] <http://download.eclipse.org/egit/updates-4.0.2/plugins/> And ... you know today's the "last day" for contributions, right? :/ (In reply to David Williams from comment #4) > (In reply to Andreas Sewe from comment #3) > > FYI, the current EGit contribution (4.0.2) [1] for RC4 apparently still > > contributes JSch 0.1.51 [2] > > > > @EGit team: Is this intentional? > > > > [1] > > <https://git.eclipse.org/c/simrel/org.eclipse.simrel.build.git/tree/egit. > > b3aggrcon?h=Mars_maintenance&id=b309f64b0f90cef6495f8769f3681907b305dcac> > > [2] <http://download.eclipse.org/egit/updates-4.0.2/plugins/> > > And ... you know today's the "last day" for contributions, right? :/ Yes, I do. That's why I pinged the EGit team on this bug again; I am not a EGit contributor myself, just the guy who noticed the two versions in the repo and filed a bug report about it. (In reply to Andreas Sewe from comment #5) > Yes, I do. That's why I pinged the EGit team on this bug again; I am not a > EGit contributor myself, just the guy who noticed the two versions in the > repo and filed a bug report about it. I looked at one of the RC4 EPP Packages (JEE) and, sure enough, both are there, as part of that package. I suspect due to the way features are specified. Hard to say what would happen at runtime ... probably the highest one would be chosen, depending on how everyone specifies their version range constraints, but I know we in the Platform specify a "wide" range as "ok" (namely "[0.1.50,1.0.0)"). I think RSE (with it's SSH Terminal support) had a stronger demand for 0.1.53 and its improved encryption methods ... and not sure how they specify theirs. Just presenting this as "data", not saying there's a problem (other than the tiny increase in size, for having two of them in the package). > Just presenting this as "data", not saying there's a problem (other than the > tiny increase in size, for having two of them in the package). Well ... might be a problem with some "update paths"? See bug 420903 comment 56. (In reply to Andreas Sewe from comment #3) > FYI, the current EGit contribution (4.0.2) [1] for RC4 apparently still > contributes JSch 0.1.51 [2] > > @EGit team: Is this intentional? > > [1] > <https://git.eclipse.org/c/simrel/org.eclipse.simrel.build.git/tree/egit. > b3aggrcon?h=Mars_maintenance&id=b309f64b0f90cef6495f8769f3681907b305dcac> > [2] <http://download.eclipse.org/egit/updates-4.0.2/plugins/> this is not intentional, we changed the JGit / EGit Mars target platform to use 0.1.53 https://git.eclipse.org/r/54117 https://git.eclipse.org/r/54119 and set the corresponding maven profile as default but unfortunately I missed that we have set maven profiles explicitly in our Hudson build jobs and missed to also explicitly select the maven profile choosing the mars target platform. Hence my local test build for 4.0.2 contained 0.1.53 but the result of the 4.0.2 release build which we run on Hudson contains 0.1.51. I fixed the build job configurations but it's too late to fix this for Mars.1 :-( We are releasing 4.1 this week and I will double check that it really contains jsch 0.1.53 before publishing the release. Just out of curiosity, if we do a respin of Mars.1, (see bug 478151) would you have ability and interest in providing a fix for Mars.1, so that only 0.1.53 was pre-req'd? (And, where that would be the only change). I'm not saying you should, necessarily, but IMHO, this problem *might* be worse than the one we are respining for, since there is no easy work around (other than "-clean" ... which in theory can have other side effects. Plus, I am not positive that having just 0.1.53 in the 'update' would 100% fix the issue reported in bug 420903 comment 56. Just thought it worth thinking about, and hearing your opinion. Thanks, I could tag 4.0.3 with jsch 0.1.53 and contribute it to the Mars.1 respin. It's a matter of a few hours. (In reply to Matthias Sohn from comment #10) > I could tag 4.0.3 with jsch 0.1.53 and contribute it to the Mars.1 respin. > It's a matter of a few hours. Ok, please do. This is just a change to included version of com.jcraft.jsch in org.eclipse.jgit, right? Assuming so, please do prepare it, and I'll include it, unless someone on Planning Council objects. Thanks. (In reply to David Williams from comment #11) > (In reply to Matthias Sohn from comment #10) > > I could tag 4.0.3 with jsch 0.1.53 and contribute it to the Mars.1 respin. > > It's a matter of a few hours. > > Ok, please do. This is just a change to included version of com.jcraft.jsch > in org.eclipse.jgit, right? yes, compared to 4.0.2 it will include JSch 0.1.53 instead of 0.1.51, no other changes except flipping version numbers > Assuming so, please do prepare it, and I'll include it, unless someone on > Planning Council objects. ok, I'll prepare it later today When you are ready, please contribute to the Mars_maintenance.1 branch, of the simrel.build repo. (And, just as well do it directly ... no plans to change Gerrit job). And, just as well update the Mars_maintenance release too, so "in sync". submitted changes for 4.0.3 and contributed to Mars.1 respin I don't know if you take into consideration the bug 477466: [regression] Remote system explorer fails to connect SSH due to JSch-0.1.53 in stops working with Eclipse Mars.1. I've reported the above bug last week, just to let you know about. |