No announcement yet.

Poor XML comparison

  • Filter
  • Time
  • Show
Clear All
new posts

  • Poor XML comparison

    I'm evaluating BC3 at the moment and the overall experience is positive.

    I've been comparing some XML files the last few days and have been disappointed with the comparison results. It's related to automatic alignment. I simply delete a single XML block but the Diff result shows two edits. The alignment is getting confused by similar looking closing tag symbols. Manually aligning sometimes works around the issue (slow), other times even that does not work, with the comparison now showing a new third difference even though blocks are identical.

    I have not attached a XML example file, however I think the issue is reproducible enough. Just grab one of your visual studio project files and delete some tags to simulate project changes.

    To work around this issue I've:
    1) Played with the alignment settings in options menu.
    2) Selected XML for the file format, which did show correct syntax highlighting but nothing else.
    3) Tried comparing using AraxisMerge and Diffuse, which both showed correct alignment and comparison.

  • #2

    When you played around with the Alignment settings, did you try the Alternate Alignment method? This is an entirely different algorithm that doesn't support closeness matching, is a bit slower, and is missing some other features, but sometimes shows results closer to other programs.

    If you could send in sample files to [email protected] it would be appreciated. We have a wide variety of XML sample cases and always appreciate more to test against as we improve/enhance this area. We do not use Visual Studio in office. If you email us, please include a link to this forum thread for our reference.
    Aaron P Scooter Software


    • #3
      Thanks Aaron, I will try to find a complex case, but if not, I'll send a simple repro for you to look at.

      To any readers following at home, the other Diff/Merge programs I mention are ones I used to use and am dissatisfied with, which is why I'm here. BC3 looks to be both well featured and good value.


      • #4
        Test case has been mailed, thank you.

        I tried to reproduce the manual alignment case but could not. I think it occurs when I do not select the first possible line to align with. In this case BC3 correctly aligns from the later point, but does not look back to notice alignment could have started earlier as corresponding lines are now identical.


        • #5

          Thanks for the sample files. I've reproduced the alignment trouble you are seeing and emailed a follow-up about adding these to a list of test cases we have. Our current algorithm won't quite align these automatically as you expect, and would need manual tweaking to push different xml nodes, begin and end sections, into place.
          Aaron P Scooter Software