Creating a Working Scirun Directory Tree
Prev C. Local SCIRun Source Code Control Using CVS Next

Creating a Working Scirun Directory Tree

This procedure assumes the latest SCIRun/BioPSE tar files have already been downloaded, and the source code has been imported, and merged if necessary, into the repository. The following steps build a working directory by extracting the Thirdparty software, DataSets, and documentation from tar files and obtaining the SCIRun/BioPSE source code from the CVS repository. This procedure also assumes, for the present, that the BioPSE and MatlabInterface packages are desired in the working directory, and no other packages have been imported into the repository. Under this model, it is expected that each developer will be writing and compiling code in his/her own working directory, independently of other developers. Periodically, all developers will check their work in (commit) to the common repository. 

  1. setenv CVSROOT /proj/cdsp/biomed2/cvsSCIRun

    or to retrieve source code via ssh: 

    setenv CVSROOT gateway.ece.neu.edu:/proj/cdsp/biomed2/cvsSCIRun
setenv CVS_RSH ssh

  2. Create a top-level SCIRun directory. This can be anything because it is a personal working directory, eg., /usr/people/scirun/SCIRun.1.6.0.lk or /export/scratch/SCIRun.1.6.0.test.

  3. cd into this directory

  4. Assuming unzipped tar files reside in /proj/cdsp/biomed2/SCIRunDownloads/vx.x.x_tars, issue the following commands (version numbers given as an example): 

    1. tar xvf \
/proj/cdsp/biomed2/SCIRunDownloads /v1.6.0_tars/Thirdparty_install.1.6.0
	    
    2. tar xvf /proj/cdsp/biomed2/SCIRunDownloads /v1.6.0_tars/DataSets.1.6.0.tar
	    
    3. tar xvf /proj/cdsp/biomed2/SCIRunDownloads /v1.6.0_tars/docs.1.6.0.tar

  5. cd into the Thirdparty directory created in the previous step, eg., 

    cd /export/scratch/SCIRun.1.6.0.test/Thirdparty_install.1.4.3.

    Note

    The Thirdparty software distributed with v1.6.0 is unchanged from v1.4.2. This is why the numbers 1.4.3 and 1.4.2 appear.

    Run the install script:

    python install INSTALLDIR [irix, Linux] [sgi, gcc] [32, 64]

    where INSTALLDIR is the top-level working directory, eg., /export/scratch/SCIRun.1.6.0.test, and brackets indicate lists of options for each argument.

    This creates a 1.4.2 directory (because the thirdparty software has not changed with v1.6.0) under INSTALLDIR containing all the necessary thirdparty software libraries.

    The previous 5 steps take care of the SCIRun installation that is independent of CVS. The remaining steps retrieve the SCIRun/BioPSE source code that has been imported to the repository, and build the SCIRun program.

  6. cd back to INSTALLDIR

  7. Create a SCIRun directory below INSTALLDIR

  8. cd into SCIRun

  9. Enter the following command to checkout the HEAD revision of the SCIRun/BioPSE source code files. Note that the period at the end of the line below is an essential part of the command.

    cvs checkout .

    This command creates and populates the src and doc directories below ./SCIRun.

  10. Create an sgi32 or linux (or appropriate other) directory below the ./SCIRun directory, and cd into this directory.

  11. Run configure as follows (eg.):

    ../src/configure
--with-thirdparty=/export/scratch/SCIRun.1.6.0.test/1.4.2/Linux/gcc-32bit  
--enable-package='BioPSE MatlabInterface'--enable-debug

    If configure does not run successfully under sgi, try setting the following environment variables and repeat the configure command.

    setenv CC cc
setenv CXX CC
setenv F77 f77

  12. Compile SCIRun/BioPSE:

    gmake -j JOBNUMBER

    where 5 is a typical value for JOBNUMBER

    SCIRun may appear to compile successfully, but run-time errors may occur if environment variables starting with LD_LIBRARY were set prior to compilation. Therefore, it is recommended to unset these environment variables when compiling and running SCIRun, e.g.,

    unsetenv LD_LIBRARY_PATH
Prev Up Next
Importing a Release Home Common CVS Commands