Guilty.. I guess we should be putting them under 0.11 in trunk.
On Tue, Oct 16, 2012 at 8:18 PM, Jonathan Coveney <[email protected]> wrote: > AFAIK (and I don't really know), I thought that if we put it in both, that > it'd go in the pig 11 section in trunk, and if not, we don't. > > Is this correct? > > Good job noticing this. > > 2012/10/16 Gianmarco De Francisci Morales <[email protected]> > >> Hi devs, >> >> I noticed there is a misalignment in CHANGES.txt between 0.11 and trunk. >> It seems some people are putting patches on top in both versions of >> the file, while other are putting changes that get into 0.11 in the >> 0.11 section of the trunk file. >> >> Let me show an example: >> >> This is 0.11 >> >> Pig Change Log >> >> Release 0.11.0 (unreleased) >> >> INCOMPATIBLE CHANGES >> PIG-1891 Enable StoreFunc to make intelligent decision based on job >> success or failure (initialcontext via gates) >> >> IMPROVEMENTS >> PIG-2947: Documentation for Rank operator (xalan via azaroth) >> >> PIG-2943: DevTests, Refactor Windows checks to use new Util.WINDOWS >> method for code health (jgordon via dvryaboy) >> >> PIG-2794: Pig test: add utils to simplify testing on Windows (jgordon via >> gates) >> >> PIG-2908: Fix unit tests to work with jdk7 (rohini via dvryaboy) >> >> PIG-2965: RANDOM should allow seed initialization for ease of testing >> (jcoveney) >> >> PIG-2964: Add helper method getJobList() to PigStats.JobGraph. Extend >> visibility of couple methods on same class (prkommireddi via >> billgraham) >> >> >> >> >> >> >> And this is trunk: >> >> Pig Change Log >> >> Trunk (unreleased changes) >> >> INCOMPATIBLE CHANGES >> >> IMPROVEMENTS >> >> PIG-2943: DevTests, Refactor Windows checks to use new Util.WINDOWS >> method for code health (jgordon via dvryaboy) >> >> PIG-2966: Test failures on CentOS 6 because MALLOC_ARENA_MAX is not >> set (cheolsoo via sms) >> >> PIG-2793: Pig test: add utils to simplify testing on Windows (jgordon via >> gates) >> >> PIG-2908: Fix unit tests to work with jdk7 (rohini via dvryaboy) >> >> OPTIMIZATIONS >> >> BUG FIXES >> >> PIG-2928: Fix e2e test failures in trunk: FilterBoolean_23/24 >> (cheolsoo via dvryaboy) >> >> Release 0.11.0 (unreleased) >> >> INCOMPATIBLE CHANGES >> PIG-1891 Enable StoreFunc to make intelligent decision based on job >> success or failure (initialcontext via gates) >> >> IMPROVEMENTS >> >> PIG-2947: Documentation for Rank operator (xalan via azaroth) >> >> PIG-2910: Add function to read schema from outout of Schema.toString() >> (initialcontext via thejas) >> >> PIG-2965: RANDOM should allow seed initialization for ease of testing >> (jcoveney) >> >> PIG-2964: Add helper method getJobList() to PigStats.JobGraph. Extend >> visibility of couple methods on same class (prkommireddi via >> billgraham) >> >> >> >> Notice how PIG-2943, PIG-2793, PIG-2908 are marked as appearing in >> trunk in trunk and in 0.11 in 0.11. >> PIG-2910 is in 0.11 in trunk but not in 0.11 (I guess it is a small >> mistake). >> >> >> So, what's the correct behavior? >> Do we mark a patch in CHANGES.txt at the earliest place it appears in >> the code (so that CHANGES.txt is consistent across releases)? >> Or do we treat the branches independently, and thus we put each patch >> always at the top? >> >> Personally, I put PIG-2947 in the 0.11 section in trunk, but I don't >> have a strong opinion on it (as long as we are consistent). >> >> Cheers, >> -- >> Gianmarco >>
