Menu
▾
▴
Tree [6cd4a1] master / History
Read Me
*Authors* --------- David Car (david.car7@gmail.com) Michael List (michael.list@gmail.com) *Requirements* -------------- gnu make Python > 2.5 To build the examples with PyF95++ you also need: gcc Fortran compiler: gfortran 4.3, ifort or pgfortran The unit tests Makefile assumes gfortran. *Notes* ------- * Version 1.5.0 - This version represents a significant departure from the 1.4.x series of the BlockIt/PyF95++ code base. The 1.5.x series has dropped the Library functionality in favor of caching parsed files using the filesystem through a new CacheManager class. To facilitate this change, all blocks are now pickleable (serializable) directly. Currently, the block parsing process in multithreaded with plans to have all phases multithreaded. There are two multithreaded implementations under the hood of PyF95++. One uses the threading module available in python 2.5 and higher. The other uses the multiprocessing module only available in python 2.6 and higher. You are strongly encouraged to upgrade to python 2.6 since the multiprocessing module is the superior choice as it will facilitate the multithreading of all phases. * Version 1.4.3 - With this version, you will need to do a `make uninstall` prior to installing this version. There has been some small directory changes and such. Also, this will force the rebuild of the STL.lib. *Documentation* --------------- The beginning of the documentation exists in the doc directory. It is a work in progress at the moment. You'll find a tex file in there and can build it with pdfLaTeX or LaTeX. *Installation* -------------- NOTE: There's been problems with the PyF95++ install failing. It's been traced to the Makefile not properly setting the PYTHONPATH environment variable prior to running PyF95++. I haven't been able to track this down, but a certain fix is to set the environment variables in the secion *Environmnet Variables* below BEFORE doing 'make install'. FYI. To install the blockit package *only* you can simply do: python setup.py install which uses the Python distutils package to install it. The default installation prefix location is in ~/slash which can be found in the setup.cfg file. If you would like a different location, then please set the 'prefix' variable accordingly, i.e. prefix=/foo/bar/slash would install the blockit package in: /foo/bar/slash/lib/python2.x/site-packages If you want to install the blockit package and the PyF95++ code, then do: make install If you have a *PREVIOUS VERSION INSTALLED*, then do: make uninstall before 'make install' to clean out any old templates and such. Also, sometimes files are renamed and this will clear out the old names. This will install PyF95++ under the PREFIX variable found in the Makefile. Currently, this is set to ~/slash. Then PyF95++ components will be installed under the following directories: $(PREFIX)/bin $(PREFIX)/lib/pyf95++ $(PREFIX)/include/pyf95++ $(PREFIX)/share/pyf95++/templates $(PREFIX)/share/pyf95++/podModules If there is a problem building the STL.lib file, please setup the environment variables listed in the section *Environment Variables* below. I've had some issues with the PYTHONPATH not being set correctly from the Makefile and it failing. I haven't been able to track this down yet. *Environment Variables* ----------------------- In order to use PyF95++ after the install, set the following environment variables: PyF95_TEMPLATE_DIR=$(PREFIX)/share/pyf95++/templates PyF95_SRC_DIR=$(PREFIX)/share/pyf95++/src PyF95_POD_DIR=$(PREFIX)/share/pyf95++/podModules PyF95_ADDON_DIR=$(PREFIX)/share/pyf95++/addons PyF95_INC_DIR=$(PREFIX)/include/pyf95++ PYTHONPATH=$(PREFIX)/lib/python2.x/site-packages where the 2.x in python2.x above is your current Python version. So if you have Python 2.6 installed on your system, then: PYTHONPATH=$(PREFIX)/lib/python2.6/site-packages *Uninstall* ----------- Simply do: make uninstall *Abstract* ---------- BlockIt is a Python framework that allows you to quickly define what blocks are within your code and parse your code into a b+-tree of blocks. At face value, this doesn't appear very useful, but with it you can build a fairly sophisticated pre-processor that allows for some pretty creative software. In particular, it has allowed me to fairly easily build a templating (i.e. generic programming) feature for the F90/95 language and from this develop a standard template library. The BlockIt module contains a few key base classes to inherit from as well as some functions to parse and manipulate a template string; create dependency graphs based on your blocks; topologically sort those graphs. The documentation of the framework is currently in the works, but to get a feel for what you can do, please take a look at the PyF95++ directory which contains PyF95++. Also, the blockit package contains the F95 package which is the module built on the BlockIt module and shows how the blocks are defined for the F90/95 language with some special block extensions, along with some key methods they should have. The PyF95++.py is the executable pre-processor which performs the templating. *Example PyF95++ Build from the PyF95++/tests directory* -------------------------------------------------------- You can build the entire set of unit tests by typing the following: make unit-tests The Makefile assumes gfortran. This will build a series of unit tests for the Standard Template Library for Fortran. If this works, you're in good shape. If you frequently pull from the git repo, you can execute this make again and it will be much faster the second time since it will use a cached version of the file if the file has not changed. This results in much faster builds.
