Automating Project Setup

The "manual" way of configuring an SVEitor project is given in the site Setting_up_an_SVEditor_Project. This takes a number of steps, and when working between multiple projects, user areas, can get a bit tedious. Fortunately this process can be automated with minimal effort.

Two files need to be created, after which the newly created project can be imported. The two files are .svproject and .project, which will typically reside at the top of your user area. A simple <perl|other_favorite_scripting_language> script is all that is required to create these flies. After you create this script, it should be as easy as running the script as part of your user area creation.

The easiest way to do this process is to create and configure a project using the GUI. When you are happy with the way the project runs the project files can be used as templates for your script (or use them to modify the attached scripts).

Creating .project

The .project file contains the following items that I will be describing.

  • Name of the project
  • List of environment variables required
  • Directories to include / exclude

Project Name

The project name can really be anything. The Project name is used in:

  • The Project Explorer view
  • Leading the path when you Open Resource

Given the above, keep the Project name short, unique, and reasonably descriptive. Often you will have multiple projects loaded into Eclipse, make it easy for yourself and others.

Example:

  • <chip_codename>_<user_sandbox> - 555_builderb_01 (assuming you are working on a 555 timer, and your name is Bob the Builder, and this is your first sandbox)

The project name is set when you go through new project creation in the following dialog:

In the .project file, this can be found near the top of the .project file:

<projectDescription>
   <name>Chip1</name>

Environment Variables

There is a decent chance that your environment uses assorted environment variables, making it easy for your scripts to work in a user sandbox independent way. At a minimum most groups will have at least some version of SANDBOX_LOCATION or PROJECT_USER_LOCATION specified. If any of these environment variables are used in your files.f's, you'll want to include these environment variables in your project.

Using the GUI, these are added using the Project Properties>Resource>Linked Resources as shown in the picture below. Eclpise includes a number of standard variables which are useful including:

  • ECLIPSE_HOME
  • PARENT_LOC
  • PROJECT_LOC - Should be the top of your user area!!!
  • WORKSPACE_LOC

In the example below, I have added a variable, CHIP which is identical to PROJECT_LOC. Chip in this example is the path to the root of my user sandbox.

Path variables can be identified in the .project file by looking for variable_list section as shown in the example below: 

   <variableList>
     <variable>
        <name>CHIP</name>
        <value>$%7BPROJECT_LOC%7D</value>
     </variable>
  </variableList>

Resource Filtering

Resource filtering is another massively useful tool, which, if used can "lighten the load" for Eclipse, by filtering out files that Eclipse will never use, and should skip while indexing.

Using the GUI, you'll set these variables using Project Properties>Resource>Resource Filters>. The example below shows me telling Eclipse to Exclude all files and folders, recursively that are named log.

NOTE: For some reason I have found you more or less have to always files&folders, recursively to get the thing to work. Driving me nuts.

The resource filters are stored in the .project file in the filteredResources section. Example of this given below. The last argument in the <arguments> section is the name of the directory to be excluded. In the example below we are ignoring anything in log and dir_to_exclude. 

  <filteredResources>
     <filter>
        <id>1390544795288</id>
        <name></name>
        <type>30</type>
        <matcher>
           <id>org.eclipse.ui.ide.multiFilter</id>
           <arguments>1.0-name-matches-false-false-log</arguments>
        </matcher>
     </filter>
     <filter>
        <id>1390544911517</id>
        <name></name>
        <type>30</type>
        <matcher>
           <id>org.eclipse.ui.ide.multiFilter</id>
           <arguments>1.0-name-matches-false-false-dir_to_exclude</arguments>
        </matcher>
     </filter>
  </filteredResources>

Creating .svproject

The file .svproject is used by SVEditor to manage it's variables.  Of these, the most important is the location of the files.f(s) which gives SVEditor a list of files in your project.

This is set via the GUI in Project Properties>System Verilog Project Properties>Argument Files.

This will appear in your .svproject file in the argFilePaths section 

  <argFilePaths>
     <argFilePath path="${project_loc}/sim/files.f"/>>
  </argFilePaths>

Using the Script

The script itself is reasonably simple to use. It is a Perl script. By the way, if you haven't already done so, install the EPIC perl editor in Eclipse, it rocks. In my example I am using a .project/.svproject generated by Eclipse 4.3.1, and SVE version 1.5.0.

Editing the Script

Follow the following steps to get the automation in progress:

  1. Download create_eclipse_setup.pl
  2. Open the file up in Eclipse
  3. Search for "Step 1" - Replace the project name with the project name of your choice
  4. Search for "Step 2" - Add any environment variables you need
  5. Search for "Step 3" - Add any directories you want exclude
  6. Search for "Step 4" - Add all files.f's that you need
  7. Run the script. On a command line, run ./create_eclipse_setup.pl
  8. Import the project:
    1. In Eclipse: File>Import>General>Existing Projects into Workspace>Next
    2. Browse to the top of your user area (where the newly created .project and .svproject reside)
  9. If it looks good, add the script to your repository, and have it run automatically when a user area is created!!!

Optional Steps

  • Using the Eclipse GUI as described in the sections above, create .project and .svproject files. The GUI will give correct syntax in the files. This can then by copy & pasted into the script, which may make certain things easier.
  • There are scripts available which will make it easy to "grab" your environment variables, making "Step 2" above less manual.