I fixed HHH-1780, may relates to this issue? ----------- Strong Liu <stliu(a)hibernate.org&gt; http://hibernate.org http://github.com/stliu On Aug 2, 2011, at 5:20 AM, Scott Marlow wrote: > http://pastebin.com/SR3jsxGp contains the call stack for a > ClassNotFoundException of org.hibernate.hql.internal.ast.HqlToken (from > https://issues.jboss.org/browse/AS7-1433). The exception message isn't > shown but the jira has that. > > I tried the supplied test case against Hibernate 4.0.0.Beta3 and don't > have the same problem. Does anyone know of a Hibernate 4.0.0.Beta4 > change that could of caused this? > > Perhaps we need to set the TCCL to the hibernate CL when constructing > the HqlBaseLexer as a workaround (the Hibernate classloader depends on > the antlr classloader module, so that should work). > > Whatever the solution is, I'd like to better understand what changed to > cause this. > > Thanks, > Scott > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev

Anyone else have any ideas of what might of changed. Perhaps we were getting lucky in Beta3 and some unrelated bug was fixed that somehow was setting the TCCL for ANTLR. On 08/01/2011 05:20 PM, Scott Marlow wrote: > http://pastebin.com/SR3jsxGp contains the call stack for a > ClassNotFoundException of org.hibernate.hql.internal.ast.HqlToken (from > https://issues.jboss.org/browse/AS7-1433). The exception message isn't > shown but the jira has that. > > I tried the supplied test case against Hibernate 4.0.0.Beta3 and don't > have the same problem. Does anyone know of a Hibernate 4.0.0.Beta4 > change that could of caused this? > > Perhaps we need to set the TCCL to the hibernate CL when constructing > the HqlBaseLexer as a workaround (the Hibernate classloader depends on > the antlr classloader module, so that should work). > > Whatever the solution is, I'd like to better understand what changed to > cause this. > > Thanks, > Scott > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

https://github.com/scottmarlow/hibernate-core/commit/07d15fb39b1ac7e28546... seems to around it, by setting the TCCL to the Hibernate classloader around the Antlr call "setTokenObjectClass( HqlToken.class.getName() )" Any objections to this change? On 08/02/2011 11:12 AM, Scott Marlow wrote: > I created HHH-6536 for this issue. > > On 08/02/2011 09:46 AM, Scott Marlow wrote: >> Anyone else have any ideas of what might of changed. Perhaps we were >> getting lucky in Beta3 and some unrelated bug was fixed that somehow was >> setting the TCCL for ANTLR. >> >> >> On 08/01/2011 05:20 PM, Scott Marlow wrote: >>> http://pastebin.com/SR3jsxGp contains the call stack for a >>> ClassNotFoundException of org.hibernate.hql.internal.ast.HqlToken (from >>> https://issues.jboss.org/browse/AS7-1433). The exception message isn't >>> shown but the jira has that. >>> >>> I tried the supplied test case against Hibernate 4.0.0.Beta3 and don't >>> have the same problem. Does anyone know of a Hibernate 4.0.0.Beta4 >>> change that could of caused this? >>> >>> Perhaps we need to set the TCCL to the hibernate CL when constructing >>> the HqlBaseLexer as a workaround (the Hibernate classloader depends on >>> the antlr classloader module, so that should work). >>> >>> Whatever the solution is, I'd like to better understand what changed to >>> cause this. >>> >>> Thanks, >>> Scott >>> _______________________________________________ >>> hibernate-dev mailing list >>> hibernate-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >> >> _______________________________________________ >> hibernate-dev mailing list >> hibernate-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/hibernate-dev > > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

On 08/02/2011 04:34 PM, Scott Marlow wrote: > I'll add a security manager check to the get/set TCCL, do the manual > test in AS7 once more and commit/merge to master. I'll hold off on using SM, since we don't seem to do that everywhere. I don't want to break consistency. ;) > > On 08/02/2011 02:10 PM, Scott Marlow wrote: >> https://github.com/scottmarlow/hibernate-core/commit/07d15fb39b1ac7e28546... >> seems to around it, by setting the TCCL to the Hibernate classloader >> around the Antlr call "setTokenObjectClass( HqlToken.class.getName() )" >> >> Any objections to this change? >> >> On 08/02/2011 11:12 AM, Scott Marlow wrote: >>> I created HHH-6536 for this issue. >>> >>> On 08/02/2011 09:46 AM, Scott Marlow wrote: >>>> Anyone else have any ideas of what might of changed. Perhaps we were >>>> getting lucky in Beta3 and some unrelated bug was fixed that somehow was >>>> setting the TCCL for ANTLR. >>>> >>>> >>>> On 08/01/2011 05:20 PM, Scott Marlow wrote: >>>>> http://pastebin.com/SR3jsxGp contains the call stack for a >>>>> ClassNotFoundException of org.hibernate.hql.internal.ast.HqlToken (from >>>>> https://issues.jboss.org/browse/AS7-1433). The exception message isn't >>>>> shown but the jira has that. >>>>> >>>>> I tried the supplied test case against Hibernate 4.0.0.Beta3 and don't >>>>> have the same problem. Does anyone know of a Hibernate 4.0.0.Beta4 >>>>> change that could of caused this? >>>>> >>>>> Perhaps we need to set the TCCL to the hibernate CL when constructing >>>>> the HqlBaseLexer as a workaround (the Hibernate classloader depends on >>>>> the antlr classloader module, so that should work). >>>>> >>>>> Whatever the solution is, I'd like to better understand what changed to >>>>> cause this. >>>>> >>>>> Thanks, >>>>> Scott >>>>> _______________________________________________ >>>>> hibernate-dev mailing list >>>>> hibernate-dev(a)lists.jboss.org >>>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>>> >>>> _______________________________________________ >>>> hibernate-dev mailing list >>>> hibernate-dev(a)lists.jboss.org >>>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >>> >>> _______________________________________________ >>> hibernate-dev mailing list >>> hibernate-dev(a)lists.jboss.org >>> https://lists.jboss.org/mailman/listinfo/hibernate-dev >> >> _______________________________________________ >> hibernate-dev mailing list >> hibernate-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/hibernate-dev > > _______________________________________________ > hibernate-dev mailing list > hibernate-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/hibernate-dev _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

On Tue, 02 Aug 2011 23:03:47 +0200, Scott Marlow&lt;smarlow(a)redhat.com&gt; wrote: > I pulled my commit but it came out wrong > (5dc04960d6ccb4f54f49c2fac83a0ccd2874fa2f merge got created). It looks > like I should of rebased the commit on my local master branch before > doing the pull request. It looks like git merged my change and Hardy's > (from earlier today). > > One thing that is worse than what I did, is rebasing after the fact (I > think). Rather than making this worse, I'll wait for feedback from > others that have been through this. No problem. This is not the first merge commit (see https://github.com/hibernate/hibernate-core/network).

Rebasing now and doing a forced push would definitely be worse now, especially since Gail committed in the meantime as well. --Hardy _______________________________________________ hibernate-dev mailing list hibernate-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev

> I just pulled another change and got another merge. I rebased my repo > (git pull --rebase upstream master) against master first but it looks > like I might need to delete my repo and recreate it (cheap way to > combine parent 2d52494b90507ba676ef + parent 72078ade4d1f293d5c49). > > https://github.com/hibernate/hibernate-core/commit/cefa777c0089eac7456d50... Hi Scott, from the network graph it looks like your commit was fine and should have been fast-forwarded, so I don't think there's something wrong with your repository. Are you using the github web ui to handle the pull request? If so, that one is always going to introduce a merge commit, so we avoid it.

In other cases if you think your local repository is not in sync anymore, you don't need to throw it away, just delete your master branch, fetch from upstream and then look at the commit id of the upstream master (can find it on github too), and create it as your local branch

for the current state assuming you're in a different branch and you deleted master: git checkout -b master cefa777c0089eac7456d I just had to do this myself since it seems my master history diverged again from upstream :( Cheers, Sanne