sorry - i did mean the 3.2.8 vote thread (not 3.3.2). On Mon, Apr 9, 2018 at 9:12 AM, Stephen Mallette <[email protected]> wrote:
> Just a quick note - we determined in the 3.3.2 vote thread: > > https://lists.apache.org/thread.html/a532196a7983e9a66b8949def5a3d4 > 2c40c5155c86f980ea08f0b853@%3Cdev.tinkerpop.apache.org%3E > > that we'd try to get gremlin-javascript deployed under "gremlin" in npm so > as not to have to re-do the whole release. Anyway, it was more complicated > than just changing the package file to do that, so ultimately, I ended up > deploying to both "gremlin-javascript" and "gremlin". Documentation was > post updated to point to "gremlin" but the READMEs and packaged docs still > point to gremlin-javascript. Ultimately we will promote "gremlin" as the > "right" one but both will work. We'll get gremlin-javascript out of here in > the next release. Wish we'd had caught all that sooner.....oh well. > > On Fri, Apr 6, 2018 at 11:00 PM, Robert Dale <[email protected]> wrote: > >> Looks good. Doesn't even show up as a forced update. >> >> Robert Dale >> >> On Fri, Apr 6, 2018 at 9:36 PM, Stephen Mallette <[email protected]> >> wrote: >> >> > i think i fixed it. i accidentally pushed a really bad version of tp33 >> that >> > made it even with master!!! >> > >> > i ended up fixing by doing: >> > >> > git branch -D tp33 >> > git checkout tags/3.3.2 >> > git checkout -b tp33 >> > ... // all the stuff to bump to 3.3.3-SNAPSHOT again >> > git push origin tp33 --force >> > >> > Thanks to Robert Dale for being available to help me with a git sanity >> > check on that. It would be great if others could take a look at the git >> > history at some point soon (ideally before we go committing on it again) >> > and give me a +1 if everything looks ok. >> > >> > sorry for the foul up....missed a stupid forward slash in one of my git >> > commands and it sent tp33 into a tailspin from there and of course I >> didn't >> > realize it until i'd pushed. bah. git provides more ways for you to >> screw >> > your repo than it does to keep it straight. have a good weekend >> everyone. >> > no more git for me until next week. jeez. >> > >> > On Fri, Apr 6, 2018 at 8:54 PM, Stephen Mallette <[email protected]> >> > wrote: >> > >> > > I think I just did something bad to git..................... >> > > >> > > Please don't push anything until I can try to sort out what happened. >> > Damn >> > > >> > > On Fri, Apr 6, 2018 at 7:51 PM, Stephen Mallette < >> [email protected]> >> > > wrote: >> > > >> > >> The vote threads are now closed up but we will wait until monday to >> > >> finalize up the actual release. Have a good weekend everyone! >> > >> >> > >> On Tue, Mar 27, 2018 at 2:04 PM, Stephen Mallette < >> [email protected] >> > > >> > >> wrote: >> > >> >> > >>> I think the second issue is resolve now with: >> > >>> >> > >>> https://github.com/apache/tinkerpop/commit/2a9e7e24f4e255cad >> > >>> 3eb300423634cbac110d74d >> > >>> >> > >>> we can thank kevin gallardo for the answer on that one. >> > >>> >> > >>> On Tue, Mar 27, 2018 at 7:39 AM, Stephen Mallette < >> > [email protected]> >> > >>> wrote: >> > >>> >> > >>>> Re-running tests now: >> > >>>> >> > >>>> https://github.com/apache/tinkerpop/commit/2e74c31c39b959263 >> > >>>> 605436555ba9cbc05eab2c5 >> > >>>> >> > >>>> also, altered the original JIRA a bit: >> > >>>> >> > >>>> https://issues.apache.org/jira/browse/TINKERPOP-1884 >> > >>>> >> > >>>> >> > >>>> >> > >>>> On Tue, Mar 27, 2018 at 7:08 AM, Robert Dale <[email protected]> >> > wrote: >> > >>>> >> > >>>>> Reverting back to netty 4.0 until there is a critical bug may be >> the >> > >>>>> best >> > >>>>> course. Spark 2.3 upgrades to netty 4.1 so perhaps that can be >> done >> > in >> > >>>>> tp34/master. >> > >>>>> >> > >>>>> Robert Dale >> > >>>>> >> > >>>>> On Tue, Mar 27, 2018 at 6:55 AM, Stephen Mallette < >> > >>>>> [email protected]> >> > >>>>> wrote: >> > >>>>> >> > >>>>> > hmm - i don't like how shifty this is all getting. I think I >> might >> > >>>>> bump >> > >>>>> > tp33 back to 4.0.x then everything will just work cleanly. >> > >>>>> > >> > >>>>> > btw, i'm also now aware of a second issue. the jackson version >> bump >> > >>>>> has >> > >>>>> > created an issue where byte[] is not serializing properly. it >> used >> > >>>>> to "just >> > >>>>> > work" but something must have changed in jackson to now require >> > that >> > >>>>> it be >> > >>>>> > a registered type. i think i need to get that resolved prior to >> > >>>>> release - >> > >>>>> > this could mean a delay. >> > >>>>> > >> > >>>>> > On Tue, Mar 27, 2018 at 6:37 AM, Stephen Mallette < >> > >>>>> [email protected]> >> > >>>>> > wrote: >> > >>>>> > >> > >>>>> > > they would have a problem..... :/ >> > >>>>> > > >> > >>>>> > > i'm thinking this is an unlikely case? Spark based traversals >> are >> > >>>>> > > analytical and thus typically executed through the console >> rather >> > >>>>> than >> > >>>>> > via >> > >>>>> > > Gremlin Server. So perhaps not a big deal breaker? What do >> people >> > >>>>> think? >> > >>>>> > > >> > >>>>> > > That said, luckily the workaround is the same as the one for >> the >> > >>>>> console. >> > >>>>> > > Simply need to remove the netty 4.1.x jar from gremlin >> server's >> > >>>>> /lib and >> > >>>>> > > replace it with 4.0.x. I just ran the Gremlin Server >> integration >> > >>>>> tests >> > >>>>> > and >> > >>>>> > > they pass with that configuration. I could document that >> process >> > >>>>> in: >> > >>>>> > > >> > >>>>> > > https://github.com/apache/tinkerpop/blob/master/gremlin- >> > >>>>> > > server/conf/gremlin-server-spark.yaml >> > >>>>> > > >> > >>>>> > > >> > >>>>> > > >> > >>>>> > > On Mon, Mar 26, 2018 at 8:57 PM, Robert Dale < >> [email protected]> >> > >>>>> wrote: >> > >>>>> > > >> > >>>>> > >> What happens if someone wants to install Spark in Gremlin >> > Server? >> > >>>>> > >> >> > >>>>> > >> Robert Dale >> > >>>>> > >> >> > >>>>> > >> On Mon, Mar 26, 2018 at 4:07 PM, Stephen Mallette < >> > >>>>> [email protected] >> > >>>>> > > >> > >>>>> > >> wrote: >> > >>>>> > >> >> > >>>>> > >> > Kuppitz pointed out that we were having trouble building >> docs >> > >>>>> on tp33 >> > >>>>> > >> and >> > >>>>> > >> > of course, it was spark related. I managed to sort out the >> > >>>>> problem and >> > >>>>> > >> we >> > >>>>> > >> > have our bump to netty 4.1.x to thank along with spark 2.2 >> > being >> > >>>>> > pinned >> > >>>>> > >> on >> > >>>>> > >> > the 4.0.x line of netty. Basically spark is completely >> > >>>>> incompatible >> > >>>>> > >> there >> > >>>>> > >> > and we get errors as a result. I've tried all manner of >> things >> > >>>>> to >> > >>>>> > >> resolve >> > >>>>> > >> > it, but only one thing has actually worked. I had to pin >> the >> > >>>>> Gremlin >> > >>>>> > >> > Console to netty 4.0.x. Luckily, netty 4.0.x on >> gremlin-driver >> > >>>>> seems >> > >>>>> > >> > compatible with 4.1.x on the gremlin server so the console >> can >> > >>>>> still >> > >>>>> > >> talk >> > >>>>> > >> > to the server as needed. Here's the PR for consideration: >> > >>>>> > >> > >> > >>>>> > >> > https://github.com/apache/tinkerpop/pull/826 >> > >>>>> > >> > >> > >>>>> > >> > if anyone has a better way to resolve this, i'm open to >> going >> > >>>>> with it >> > >>>>> > - >> > >>>>> > >> in >> > >>>>> > >> > the mean time, if committers could review/vote, that would >> be >> > >>>>> good. >> > >>>>> > >> > >> > >>>>> > >> > >> > >>>>> > >> > >> > >>>>> > >> > On Sat, Mar 24, 2018 at 4:09 PM, Stephen Mallette < >> > >>>>> > [email protected] >> > >>>>> > >> > >> > >>>>> > >> > wrote: >> > >>>>> > >> > >> > >>>>> > >> > > Just a reminder that code freeze is in effect for the >> next >> > >>>>> week. >> > >>>>> > >> Please >> > >>>>> > >> > > post issues related to release of 3.2.8/3.3.2 on this >> thread >> > >>>>> as >> > >>>>> > >> needed. >> > >>>>> > >> > > >> > >>>>> > >> > >> > >>>>> > >> >> > >>>>> > > >> > >>>>> > > >> > >>>>> > >> > >>>>> >> > >>>> >> > >>>> >> > >>> >> > >> >> > > >> > >> > >
