Loading ...
Sorry, an error occurred while loading the content.
 

Another run with doxygen

Expand Messages
  • Karsten Wolf
    Hi kernel-folks, for those interested in a kernel-innards documentation: I took some more runs with doxygen http://www.doxygen.org/ on the kernel. After 15
    Message 1 of 5 , Jun 27, 2006
      Hi kernel-folks,

      for those interested in a kernel-innards documentation:


      I took some more runs with doxygen http://www.doxygen.org/ on the
      kernel.

      After 15 days worth of a G5/2.7MHz cpu time (spent 99.9% spent in
      producing dependency graphs) it produced some 40000 files taking
      400MB of disc space... the tgz file is still 190MB.

      If someone wants to host it mail me off-list.

      A version without the call-graphs is still 70/16 MB, the
      corresponding (& hyperlinked) pdf is 51/17MB.

      The archive files can't be hosted on yahoo since there is a 20MB
      total limit.


      -karsten


      P.S.: To those who know doxygen: the big version had a
      "MAX_DOT_GRAPH_DEPTH = 1000"
    • Scott Lawton
      I realize this thread is old but I just had a thought. The doxygen results are too big to post to a public web site (at least without lots of work). Instead,
      Message 2 of 5 , Jul 7, 2006
        I realize this thread is old but I just had a thought. The doxygen results are too big to post to a public web site (at least without lots of work). Instead, when someone uses one or more pages (especially in an area that others are likely to touch too), perhaps it would be useful for them to post those to the wiki (or wherever is convenient).

        Just a thought. And, an excuse to thank Karsten for the effort and computer time!

        Also, I wonder if anyone has tried to run doxygen on UserTalk code. I imagine it would require exporting to text files that maintained the table structure as folders. It's not a project I'm likely to tackle, though I'd be happy to test the result....
        --

        cheers,

        Scott
      • Seth Dillingham
        ... I m confused. Isn t that exactly what we did? Seth
        Message 3 of 5 , Jul 7, 2006
          On 7/7/06, Scott Lawton said:

          > The doxygen results are too big to post to a public web site (at least
          > without lots of work).

          I'm confused. Isn't that exactly what we did?

          <http://frontier-dox.macrobyte.net/>

          Seth
        • Karsten Wolf
          Hi Scott, ... I don t know if I understand you correctly; the doxygen html output is static and ca. 600MB in size. The pages, which are now hosted by Seth,
          Message 4 of 5 , Jul 7, 2006
            Hi Scott,


            Am 07.07.2006 um 19:57 schrieb Scott Lawton:

            > I realize this thread is old but I just had a thought. The doxygen
            > results are too big to post to a public web site (at least without
            > lots of work). Instead, when someone uses one or more pages
            > (especially in an area that others are likely to touch too),
            > perhaps it would be useful for them to post those to the wiki (or
            > wherever is convenient).

            I don't know if I understand you correctly; the doxygen html output
            is static and ca. 600MB in size. The pages, which are now hosted by
            Seth, become inaccurate with every svn commit to the main trunk. They
            reflect the main-trunk ca. mid-may. They can be useful for fast
            finding relationships in the sources (who is called by whom, which
            struct element is accessed by whom etc.)

            This one was the big one. If there's a demand, these pages can be
            kept actual; but not in that resolution. Future versions will be
            restricted to a call-depth of 5 or 6 for call-graphs; this reduces
            cpu time to ca. 6h on 876MHz G4.

            > Just a thought. And, an excuse to thank Karsten for the effort and
            > computer time!

            Honestly, it wasn't my computer that ran for 15 days ;-)


            > Also, I wonder if anyone has tried to run doxygen on UserTalk
            > code. I imagine it would require exporting to text files that
            > maintained the table structure as folders. It's not a project I'm
            > likely to tackle, though I'd be happy to test the result....


            doxygen parses the source files. AFAIK there is no parser in doxygen
            for UserTalk.



            -karsten
          • Scott Lawton
            ... Ack; sorry. That msg wasn t in the same thread and somehow I missed it. So, NEVER MIND. -- cheers, Scott
            Message 5 of 5 , Jul 7, 2006
              >I'm confused. Isn't that exactly what we did?
              >
              ><<http://frontier-dox.macrobyte.net/>http://frontier-dox.macrobyte.net/>

              Ack; sorry. That msg wasn't in the same thread and somehow I missed it.

              So, NEVER MIND.
              --

              cheers,

              Scott
            Your message has been successfully submitted and would be delivered to recipients shortly.