Looking for a procedure to compare multiple versions of multiple robots

    Hello,

    I'm a teacher and I have a hard time managing the various version of both virtual and real robots.

    I have 2 educational cells with ER-4iA (similar to LR Mate 200iD/4S) with R30iB Plus controller.

    On Roboguide, I maintain 2 different workcells for each robot. A basic version with DIO control and a more advanced version with EIP control. So eight versions total for the two virtual and two real robots.

    The tedious part is making sure that whatever modifications I make propagates to the others versions, especially for the DCS parameters, as I'm currently learning how to protect the robots from the students (!).

    So far, I managed to get all the ASCII files from the virtual robots (with ftp) and the real robots (with OneRobotics Backup Tool). When I batch compare all the file from a matching pair of robots (real and virtual), most files are identical except for the following ones that never match:

    • dcspos.va
    • diocfgsv.va
    • mixlogic.va
    • sycldint.va
    • symotn.va
    • synosave.va
    • syshost.va
    • sysmacro.va
    • system.va

    I understand that some files might be monitoring something, therefore always changing ? Like symotn.va ? If so, I can exclude them from the comparison. But for the DCS parameter, I don't want to miss something while comparing. Somebody know why DCSPOS, for example, is not matching even right after a transfer ?


    Thank you

  • 95devils May 28, 2024 at 8:30 PM

    Approved the thread.

    Have you looked into the actual comparison of files, instead of just the result? If you could see which variables differ between the two backup-ed versions, you might be able to deduce what information the variable contains. If it's merely some kind of 'last accessed' time, you could exclude it from your comparison for example. You might have to write custom comparison scripts for specific types of files, but this way you can filter (some of) the irrelevant variables from the relevant portions of the files.

    I like to use WinMerge for this, although Beyond Compare, UltraCompare, and KDiff are all options. Basically, I direct the tool to the two directories I want to compare, and it shows me a list of every file that doesn't match. Then selecting each file will open a line-by-line comparison for that file pair, highlighting not just which lines are different, but which parts of those lines.

    It's not infallible. For one thing, it only works on ASCII files, but since those are the only ones you can actually read, that's not a big loss. The bigger problem is that sometimes, certain text differences partway through a file can throw off the comparison for the rest of the file entirely, so you have to be aware of that. But I still find it extremely useful day-to-day, so much so that I have it pinned to my taskbar.

    Another thing to be aware of: many files will have "cruft" differences, like date/time stamps, that will often be different, but that difference has no actual effect on anything. So you have to keep that in mind as well. If your programs write values to Registers, then you'll have differences between your POSREG or NUMREG files that are actually harmless.


    Thanks guys, WinMerge if so much better. I was using fc (file compare) on the command line, hoping to automate the process. But I was under the impression that those files were config files. I need to rule out what is dynamic.

    I'm taking a look at WinMerge and, you're right hoitzing, i'll have to dig deeper into the comparison.

    Can I consider what is in DRAM as dynamic ? Seems to me that the controller would not put static parameter in volatile memory.

Advertising from our partners