Hi Christian, I ll try to make the fix ASAP. Also, I ll be glad if you could take part in code reviews for Apache Wave, so we could handle such issues before they become part of the source code.
On Fri, Nov 25, 2011 at 11:01 PM, Christian Ohler (Created) (JIRA) < j...@apache.org> wrote: > Bad dependency from org.waveprotocol.wave to org.waveprotocol.box > ----------------------------------------------------------------- > > Key: WAVE-306 > URL: https://issues.apache.org/jira/browse/WAVE-306 > Project: Wave > Issue Type: Bug > Components: Web Client > Reporter: Christian Ohler > Assignee: Yuri Zelikov > > > Hi Yuri, > your change > http://svn.apache.org/viewvc?view=revision&revision=1187363introduced a > reverse dependency that breaks walkaround. I believe > org.waveprotocol.wave is not supposed to depend on org.waveprotocol.box - > the dependency is supposed to go the other way only. > > Specifically, http://codereview.waveprotocol.org/616002/diff/2003/1033made > EditToolbar depend on HistorySupport. This breaks walkaround because > it makes it impossible to pull in undercurrent without also pulling in a > GWT entry point that tries to initialize socket.io. Since walkaround > doesn't use socket.io, this initialization fails. (I studied this a > while ago but didn't get around to reporting a bug back then, I hope my > recollection is not too far off.) > > Walkaround's get-third-party-deps script works around this by specifically > downloading revision 1187362 of Apache Wave, but that's not a good > long-term solution. > > Maybe HistorySupport.waveRefFromHistoryToken should be moved to > GwtWaverefEncoder? Or do you see a better way? > > > -- > This message is automatically generated by JIRA. > If you think it was sent incorrectly, please contact your JIRA > administrators: > https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa > For more information on JIRA, see: http://www.atlassian.com/software/jira > > >
