I need to be able to see the difference between the current working file and the previous commit as I am editing the file. The only history plugins I can find either use pop-ups (which disappear when I start typing) or a dedicated diff-view which takes up too much space on my small monitor.

  • OracOP
    link
    fedilink
    English
    arrow-up
    3
    ·
    edit-2
    5 months ago

    I tried that, it has different ways of looking at a diff/history, but none where you can just have it open while still working (unless you do split I suppose, but that is not practical as I need to go through hundreds of files. But perhaps it helps if I describe the problem. Maybe there is another solution I am not thinking of: I have almost 2000 files that all have had 1 change (so there is an initial commit and a recent version). What I need to do is go through each file and add the changes made to the file as a comment. So what I would like is: When I open a file, see immediately which lines were changed., and then be able to see what it was that changed so I can either copy/paste or retype it as a comment. Each change I review and I either add a #ACCEPTED, #DENIED or a #REVIEW to the comment as well.

    So having to switch to different views and opening/closing tabs for each file is something I would like to avoid. Just open the file, see what I need, type and close the file again.

    • Lodra@programming.devM
      link
      fedilink
      English
      arrow-up
      1
      ·
      5 months ago

      Can you give an example of the change and your comment for a single file?

      2000 is a lot of files. If you need to do this often, it might be best to build a small cli tool to make this easier or fully automated. It really depends on how you choose to create those comments.

      • OracOP
        link
        fedilink
        English
        arrow-up
        2
        ·
        5 months ago

        They are XML files and changes would be things like added or removed elements or changed attribute values. So e.g.: Version 1:

        [...]
        <Dynamo vendor="Siemens" max-rotation=12>
        [...]
        <Light colour="white">
        [...]
        

        Version 2:

        [...]
        <Dynamo vendor="Siemens" max-rotation=16>
        [...]
        

        The comments would look like:

        [...]
        <!-- %ATTCHANGE :: max-rotation=12 :: #ACCEPTED see report 12.24/2 -->
        <Dynamo vendor="Siemens" max-rotation=16>
        [...]
        <!-- %REM :: <Light colour="white"> :: #ACCEPTED see report 12.24/4 -->
        [...]
        

        I make the changes while going through the report (hardcopy only) which is basically an engineering report that outlines changes made to a factory (12.24/4 means chapter 12, section 24 item 4 which would list the old and new device/machine or the changes made etc. in a technical manner). It’s manual work, which is fine (I get paid by the hour). Just want it to be a bit more streamlined.

        • Lodra@programming.devM
          link
          fedilink
          English
          arrow-up
          1
          ·
          5 months ago

          Well this is definitely a bit odd. Send like those comments should be included with the original change. Or maybe they shouldn’t be recorded as comments at all and they should be a git commit message. And where does that hard copy report come from? Anyways, back to your actual question.

          At this point, I’m still suggesting a tiny utility to assist with adding the comments. It looks like %ATTCHANGE and %REM are part of a very sort list of possible values. If so, a little cli tool can definitely help there. It could also handle the general comment structure and the changed value easily. Do these comments ever include something besides #ACCEPTED?

          The tricky part is the ‘12.24/4’. It sounds like you go through the report and then find the files/lines to create these comments. Is that right? It would be tricky to code a cli tool for doing that because you need to jump around between files.

          Last note for now, some simple git commands can definitely help you here. You could easily generate a list of changed files and lines. Another could show you the changed text. For any given change. Etc.

          • OracOP
            link
            fedilink
            English
            arrow-up
            1
            ·
            edit-2
            5 months ago

            I am no git- expert; I mostly just branch and commit, that’s the extent of my knowledge. Can you point me into the direction of some information regarding this (specific commands I could use?)

            And ‘Odd’ is an understatement ;P these files are fairly old (I oldest was made in 1999 and the last version was from 2002) and the client was lucky they found the initial set and the last version on separate backups (floppies !). What we initially did was commit the initial version and then overwrite them with the latest version and commit again. So we now have a git history of the difference between the originals and the revisions. My job is to check if the final version is correct e.i either conform the work specification of the time (report), or any of the revisions (as-built) since then. And then create a comment that can be read by both the operator console and the archive-system. My best guess is someone’s nephew created this program because why anyone would use XML for this is beyond me, although I guess in 1999 this was state-of-the-art? And no one wants to touch that codebase with a 5ft pole, lol. Anyway, the comments are parsed in a specific and precise way.

            EDIT: As for your question about #ACCEPTED. The rules are “simple” each comment should start with either %ATTCHANGE, %REM, %ADD or %NOTE then 2 colons, then the old element (complete) or the ols attribute (seperated) or the note, then again 2 colons and last either #ACCEPTED, #DENIED or #REVIEW. Reveiw is followed by any text explaining why it needs reviewed, and denied and accepted are followed by either “see report xxxxx” or “as-built xxxxx” where xxxx is some reference.

            • Lodra@programming.devM
              link
              fedilink
              English
              arrow-up
              2
              ·
              edit-2
              5 months ago

              I’m on my phone so I can’t really test these out. But this should give you a few methods for printing the list of files changed.

              Here’s a similar page to show the changes for a specific file. I expect that git blame will be similarly useful.

              A few others that may help you get started as well:

              git —help
              git rev-parse
              git log
              git history
              
              • OracOP
                link
                fedilink
                English
                arrow-up
                2
                ·
                5 months ago

                thanks so much for your help.

                • Lodra@programming.devM
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  5 months ago

                  My pleasure. Thanks for posting! And feel free to ask more questions as you tackle this one.

      • thesmokingman@programming.dev
        link
        fedilink
        English
        arrow-up
        1
        ·
        5 months ago

        That’s honestly what I’d recommend. I think a quick way to do it would be to create patches from the diffs, run the patches through awk, and apply the patches.