# HG changeset patch # User Michael Guravage # Date 2010-10-15 19:27:29 # Node ID ec439fed4af87605b8c99c910764cc5c1cea274e # Parent ffb1f378901c68444f2433f3f02bd1fcaabf1dc6 Synchronise leaf movie frame and cell export data timestep references. -- user: Michael Guravage branch 'default' changed doc/v1.html changed doc/v1.pdf changed doc/v1.rst changed src/ChangeLog changed src/VirtualLeaf.cpp changed src/canvas.cpp changed src/canvas.h changed src/mainbase.h changed src/mesh.h diff --git a/doc/v1.html b/doc/v1.html --- a/doc/v1.html +++ b/doc/v1.html @@ -556,7 +556,7 @@ LanguageError: IOError: [Errno 2] No tra
  • make install
  • cd ..
  • wget http://www.riverbankcomputing.com/static/Downloads/PyQt4/PyQt-x11-gpl-4.7.7.tar.gz
  • -
  • tar xzf PyQt-x11-gpl-4.7.7
  • +
  • tar xzf PyQt-x11-gpl-4.7.7.tar.gz
  • cd PyQt-x11-gpl-4.7.7
  • python configure.py
  • make
  • @@ -594,6 +594,80 @@ still be justified - though they become introduce the new schema changes.

    +
    +

    Wednesday 13 October 2010

    +

    Tweaked .bashrc and virtualenvwrapper.sh to facilitate +virtualenv and virtualenvwrapper - many unbound variables!

    +

    Pushed recent changes to repository from where Roeland can retrieve +and test them.

    +

    OK. RhodeCode RC4 works out of the box, excepting cutehg and PyQt4.

    +
    +
    +

    Friday 15 October 2010

    +
    +

    Rhodecode

    +

    Installing RhodeCode locally on the nhypnos is a real pain. In +addition to the previous list of dependencies, for the caterpie, add +these for the nhypnos:

    +
      +
    1. Python
    2. +
    3. Bison (for sip)
    4. +
    5. Flex (for sip)
    6. +
    7. m4 (for Bison)
    8. +
    9. Qt (for PyQt)
    10. +
    11. GLIBCXX_3.4.9 (for Qt) - I draw the line at (re)installing GCC!
    12. +
    +

    OK. Forget sip bison, flex, m4, Qt and the rest. Let's install just +the minimum.

    +

    First, install sqlite; we'll see in a minute that Python requires +it. The sources are available at: http://www.sqlite.org/download.html.

    +

    Second, install Python. My initial attempt complained it was missing +the following bits and pieces:

    +
    +Python build finished, but the necessary bits to build these modules were not found:
    +_bsddb             _sqlite3           _ssl
    +_tkinter           bsddb185           dl
    +imageop            readline           sunaudiodev
    +To find the necessary bits, look in setup.py in detect_modules() for the module's name.
    +
    +

    So install sqlite first. And, since Python's configure script doesn't +have an option for it, I had to tweak setup.py to enumerate the +path where it would find sqlite;

    +

    Third, install RhodeCode. I first installed it using easy_install, but +while the result worked, the changelog view did not display the +branch/merge graphics. I then cloned the code from its bitbucket repository:

    +
    +clone https://litsol@bitbucket.org/marcinkuzminski/rhodecode
    +
    +

    Replacing the egg installed by easy_install by this code does display +the branch/merge graphics.

    +

    Fourth, setup and serve rhodecode. Following the instructions posted +online at: http://packages.python.org/RhodeCode/setup.html#setup works as advertised:

    +
    +paster make-config RhodeCode production.ini
    +paster setup-app production.ini
    +paster serve production.ini
    +
    +

    All that's left is an apache rewrite rule directing +virtualleaf.project.cwi.nl/repository to port 5000, or whaterver we +change it to, on the nhypnos.

    +
    +
    +

    The Virtual Leaf

    +

    Added incrementIterations(), getIterations() and int +iterations to mesh.h. incrementIterations() is called in TIMESTEP in +VirtualLeaf.cpp. The actual counter, iterations is initalized to +zero in the Mesh class constructor.

    +

    The iterations are inquired in two places: for the frame count in +MainBase::Plot in VirtualLeaf.cpp` and for the export cell data +interval in ``Main::TimeStepWrap in canvas.cpp.

    +

    The questions remain: Should we use this count, and if yes, where +should it be incremented? The previous count was incremented in +Main::TimeStepWrap. Or should we simply use +(int)mesh.getTime()? Since this is not an ordinal count, the +results look strange.

    +
    +