3 Replies Latest reply on Nov 8, 2006 12:55 PM by HKabaker

    Generated Duplicate Files

    Brian Gilbert
      We are maintaining a very large WebHelp project (over 1000 files in the WebHelp folder) and keeping the files in ClearCase. Everytime the project is updated all of the files are treated as though they have changed (whether they have or not). It takes a couple of hours to check the files out and into ClearCase. What I'm wondering is what files are actually updated (other than the specifically affected HTML files). Are all files in the whdata, whgdata, whxdata folders updated, only HTM files, JavaScript? All JavaScript files in the main project directory? The CAB file? XML?

      Any help would be greatly appreciated.
        • 1. Re: Generated Duplicate Files
          MergeThis Level 4
          Why are you maintaining output files in a source control product? No matter what the IT peeps say, that's a waste of time and network resources.

          If you're generating output to the default SSL directory, DON'T! You can more easily maintain the entire project directory without fooling with the output files, by generating to a completely different directory.

          As to RH output, the RH Generate function (generating output to your local machine) only replaces changed files, whereas the RH Publish function (publishing output to a server) only replaces changed files.

          By the way, doesn't ClearCase have any option to only Check In saved files? If not, get another source control product.


          Good luck,
          Leon
          • 2. Re: Generated Duplicate Files
            Brian Gilbert Level 1
            We're maintaining output files in source control because that's what we were told to do . Seriously, it feeds the testing environment and so needs to be there.

            We don't use the default !SSL! directory. We publish to a different directory entirely, so I'm glad we're following that advice.

            The RH output (Generate) republishes all files (at least, all files have a new date and time stamp and ClearCase thinks that all files have hijacked (changed)). I have checked the WebHelp Publish settings and see that the Republish All check box is not selected. I don't know why all the files have the current date on them if the RH Generate function (which is what we use) is only supposed to replace changed files.
            • 3. Re: Generated Duplicate Files
              HKabaker Level 2
              When RH finishes publishing it opens a window reporting how many files were published and listing them. You can save the list to a text file and/or print it.

              Have you compared the source code from the old and new versions and verified that RH republished a file that didn't change?

              A republished file is different in some way from the one in the target directory.

              In the whxxxx directories, RH should republish only the ones that changed.
              Search and Index terms are recompiled, but often some output files are untouched, so RH doesn't republish those.

              RH revises topic.htm files in ways that are not immediately apparent. For example, each topic refers to its TOC label, including its place in the book/sub-book structure, so a TOC revision creates an updated file even though you left the displayable contents as they were.

              Have you tried deleting the single-source output setup, which might have been corrupted, and creating a new one?

              Harvey