Question

I’m using cvs (not by choice) and I refactored a project by creating a new package hierarchy in a v2 branch. Now i need to merge the changes that happened in branch v1 into branch v2. But because the files’ path changed i can’t do a normal compare. Is there some way to do this in eclipse (or in any other tool)? i know in general, there’s no way to know for sure if the files have the same origin just by using the filename; this is a non-issue for me since all the filenames are unique in my case. I know I can go and manually select the two files and do a diff but i have too many classes to make this feasible. I am looking for an automatic pairing of the files just like when a file has changed where the tool detects the diff automatically

Does anyone have any other suggestions to do such a merge in a crappy VCS?

Update:
Some explanation of the changes I made might make it clearer

v1
Folder1
   file1
   file2
   file3
   file4
   file5

---

v2
FolderA
   file1
   FolderA1
      file2
      FolderA11
          file3
   FolderA2
      file4
FolderB
   file5

So, as can be seen, there’s no pattern in the movements. It is just a completely new hierarchy of what was before a flat hierarchy. That means i can’t do a folder diff. What I want is something that pairs v1:folder1/file4 with v2:FolderA/FolderA2/file4 for example so that i can just see the change in content without caring for the fact that the file was moved. In other words, a path-less diff that just uses the filename to match files

Answer 1

Try this shell script snippet:

ls -R /path/to/source1 | grep '.java\|.js\|.css' | grep -v ':$' >allfiles.txt
ls -R /path/to/source2 | grep '.java\|.js\|.css' | grep -v ':$' >>allfiles.txt
cat allfiles.txt | sort | uniq -d > same_files.txt
cat allfiles.txt | sort | uniq -u > different_files.txt

# This will compare the same files one-by-one
for i in `cat same_files.txt`; do 
    FILE1=`find /path/to/source1 -name $i`; 
    FILE2=`find /path/to/source2 -name $i`;
    diff -q $FILE1 $FILE2 
done

The trick is the uniq command, and its special switches. -d leaves only the duplicates, -u sorts out the duplicates.

To run on Windows, use Cygwin: http://www.cygwin.com/