I commit in 1.4, svn update, switch to my trunk directory, svn update, and then svn merge relative.path.to.1.4
So it's pretty easy command line. Just make sure you commit the directories, those hold the merge information John On Mon, Mar 5, 2012 at 11:27 AM, Keith Turner <[email protected]> wrote: > As we have been making bug fixes in the 1.4 merge branch, we have been > using svn merge to put those changes into trunk. I use eclipse to do > this. I have done it from the command line before, but do not > remember the syntax. > > Sometimes there will be a change in 1.4 that you do not want to merge > to 1.5. In this case we execute the merge, but revert the changes to > the file and check in the merge metadata. This way when someone > executes a merge in the future it will not pull those changes. I am > not sure if there is a better way to do this in svn. > > On Mon, Mar 5, 2012 at 11:10 AM, David Medinets > <[email protected]> wrote: > > That is an excellent point for which I have no great answer. I can > > avoid making changes to the 1.5 code until the 1.4 code is code > > complete... but then why have a branch at all? Let's talk about how > > the merge happens - which I am ignorant of. Does it happen file by > > file or automated with conflicts called out? > > > > On Mon, Mar 5, 2012 at 10:55 AM, Keith Turner <[email protected]> wrote: > >> I am not opposed to this change, just had a comment. One thing to > >> consider is merging bug fixes from 1.4. The more trunk is changed, > >> the more likely bug fixes in 1.4 will not merge cleanly. This > >> increases the chance that something will be lost in translation. > >> > >> Keith > >> > >> On Sun, Mar 4, 2012 at 10:52 PM, David Medinets > >> <[email protected]> wrote: > >>> In core/src/main/java/org/apache/accumulo/core/data/Mutation.java, I > >>> see the following code: > >>> > >>> > >>> private List<byte[]> values; > >>> private int cachedValLens = -1; > >>> > >>> long getValueLengths() { > >>> if (values == null) > >>> return 0; > >>> > >>> if (cachedValLens == -1) { > >>> int tmpCVL = 0; > >>> for (byte val[] : values) > >>> tmpCVL += val.length; > >>> > >>> cachedValLens = tmpCVL; > >>> } > >>> > >>> return cachedValLens; > >>> > >>> } > >>> > >>> PMD is suggesting that the line: > >>> > >>> for (byte val[] : values) > >>> > >>> should be > >>> > >>> for (byte[] val : values) > >>> > >>> Is this a useful change? >
