Propose including GEODE-7178 in 1.10

2019-09-12 Thread Michael Oleske
Hi Geode Devs! I'd like to propose including the fix for GEODE-7178. This resolves an issue that Ivan (https://markmail.org/message/dwwac42xmpo4xb2e) ran into in 1.10 RC1. SHA: 91176d61df64bf1390cdba7b1cdc2b40cdfaba3a Link to GitHub: https://github.com/apache/geode/commit/91176d61df64bf1390cdba7

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Anthony Baker
+1 yes please! > On Sep 12, 2019, at 10:11 AM, Michael Oleske wrote: > > Hi Geode Devs! > > I'd like to propose including the fix for GEODE-7178. This resolves an > issue that Ivan (https://markmail.org/message/dwwac42xmpo4xb2e) ran into in > 1.10 RC1. > > SHA: 91176d61df64bf1390cdba7b1cdc2b4

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Dick Cavender
Hi Michael, thank you for bringing your concern and fixing this issue. Geode's release process dictates a time-based schedule < https://cwiki.apache.org/confluence/display/GEODE/Release+Schedule> to cut release branches. The “critical fixes” rule does allow critical fixes to be brought to the rel

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Michael Oleske
Here is the Pull Request for the cherry pick as requested https://github.com/apache/geode/pull/4049 -michael On Thu, Sep 12, 2019 at 10:28 AM Dick Cavender wrote: > Hi Michael, thank you for bringing your concern and fixing this issue. > > Geode's release process dictates a time-based schedule

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Dan Smith
+1 for getting this in 1.10. I am curious though - is the native client behaving like an older versions of the java client, or is this totally unique behavior for the native client? Is there some integration test that we are missing here? -Dan On Thu, Sep 12, 2019 at 11:52 AM Michael Oleske wro

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Michael Oleske
The native client does behave as an old Java client (ordinal 45). I have written a story (https://issues.apache.org/jira/browse/GEODE-7190) to have Native Client updated. -michael On Thu, Sep 12, 2019 at 2:35 PM Dan Smith wrote: > +1 for getting this in 1.10. > > I am curious though - is the n

Re: Propose including GEODE-7178 in 1.10

2019-09-12 Thread Anthony Baker
My understanding is that this portion of the protocol is determined by instanceof checks, not the ordinal version. The messages from the java client went through a different code path than messages from the native client. So java clients using ordinal 45 still work (that’s why our backwards c