[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ecf-dev] Revert repacking of xpp.jar?

I just did a smoke test using Eclipse RC1, ECF most recent release build (done Wed of last week).   I'm able to connect to my google talk account, so I think that things are OK as they are/were for Wed's build (no regression afaict).

But any other testing to make sure of no regression would be appreciated.

Scott

On 5/15/2015 10:46 AM, Scott Lewis wrote:
On 5/15/2015 10:29 AM, Wim Jongman wrote:

I have no knowledge of how this should work.


You are currently ahead of everyone else since you at least know what changed in the org.jivesoftware.smack project.

What I would do would be to browse the org.jivesoftware.smack source, and find any static or dynamic references to xpp.jar classes (i.e. package org.xmlpull.*).  If the Smack classes found are 'low-level' (likely to be used for everything...like parsing the packets), and verify the tests in C-HEAD-xmpp.feature project run (they need an xmpp server but I think they are being run) then everything is ok.   

Also, if xmlpull is low-level and a smoke test of the xmpp connection to (e.g.) gmail works, then things are probably OK as well.    I will smoke test the release build via a smoke test against google talk later today/Friday, but please go ahead and a) check smack source for xpp references (low-level or not); and b) check/verify that the tests in C-HEAD-xmpp.feature are being successfully run.

Scott


Sent from Blue Mail

On May 15, 2015, at 19:21, Scott Lewis <slewis@xxxxxxxxxxxxx> wrote:
On 5/15/2015 8:33 AM, Wim Jongman wrote:
Hi, I have repacked xpp.jar to see of this resolved the problem. It was repacked from a signed jar to an unsigned jar. Basically I unpacked the jar and repacked it. This lead to a significant smaller jar. However, this did not resolve the problem BUT I did not revert that change. After that I committed the real solution which was adding some line in the build.properties file. I wonder if we should revert the repackaged xpp.jar to the original version to avoid any regression? How likely is it that the tests take care of this case?
I would hope that the xmpp tests...in C-HEAD-xmpp.feature job on Jenkins would test this. But to be direct I'm not completely clear on the use cases (in Jive's library) that lead to using/loading xpp. My belief is that it's probably used for everything...xml parsing of xmpp messages...but I'm not completely certain of that. I had intended to ask you if you tested things explicitly with the changes to xpp.jar, but didn't ask yet. I would recommend that you at least convince yourself that the C-HEAD-xmpp.feature job does at least use/test the xpp.jar usage...and that no regression has taken place. Scott
ecf-dev mailing list ecf-dev@xxxxxxxxxxx To change your delivery options, retrieve your password, or unsubscribe from this list, visit https://dev.eclipse.org/mailman/listinfo/ecf-dev


_______________________________________________
ecf-dev mailing list
ecf-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/ecf-dev



_______________________________________________
ecf-dev mailing list
ecf-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/ecf-dev