TIP is a tiny imperative programming language aimed at teaching the fundamental concepts of static program analysis. This code accompanies the lecture notes on Static Program Analysis.

Prerequisites:

• Java 8
• Scala 2.12
• SBT

We suggest you to use IntelliJ for working on TIP, but all the following options are viable.

• Install the Scala and the SBT plugins in IntelliJ.
• Clone this repository (File -> New -> Project from Version Control -> GitHub). Important: due to a bug in the SBT plugin, it is important that you choose a directory name without uppercase letters, e.g. tip (otherwise, the following step will fail).
• IntelliJ should then detect an SBT project. Click 'Import SBT project' and follow the instructions. (If this is your first Scala project, you will need to setup the Scala SDK.)
• Right-click on Tip.scala in src/tip, then select "Run 'Tip'". To supply arguments, use Run... -> Edit Configurations in the Run menu.

If you clone the repository using git instead of IntelliJ, you will need to import the project from the SBT model (File -> New -> Project from Existing Sources). Since the .idea folder is then regenerated from scrach, in order to keep the inspection profiles you need to checkout the .idea/inspectionProfiles folder and the .idea/codeStyleSettings.xml file from the original repo, leaving untouched the other generated folders.

Important: if you experience spurious type errors reported by IntelliJ for code involving Scala implicits, try disabling type-aware highlighting by clicking on the small [T] icon on the bottom right corner of the window.

If your IntelliJ has high-CPU and high-memory peaks while editing, the following tweaks might be useful:

• Disable type-aware highlighting by clicking on the small [T] icon on the bottom right corner of the window.
• Go to Help -> Edit Custom VM Options and increase the JVM memory at least with the following values:

-Xms500m -Xmx1500m

• If still nothing works, try File -> Power Save Mode.

IntelliJ offers an option to optimize imports upon commit. We suggest not to use that feature, as it may remove needed imports thereby breaking compilation.

• Check that you have installed the scala-plugin for Eclipse.
• To run TIP from within Eclipse, feed the arguments into the Run Arguments dialog.

A wrapper command tip (tip.bat for Windows) is provided to compile and run TIP with the given arguments.

Example:

./tip -run examples/fib.tip

To build:

sbt compile

Usage:

tip <options> <source> [out]

where <source> can be a file or a directory containing .tip files and [out] is an output directory (default: ./out).

To see the possible options, run tip without options. Option -verbose is recommended when developing and testing analyses.

The main function Tip.scala emits control flow graphs and analysis results as ".dot" files that can be processed by Graphviz to produce images, for example using the Graphviz dot command-line tool:

dot -O -Tpng out/example.tip__sign.dot

Some analyses require the programs use restricted subsets of TIP. The following kinds of normalization can be performed automatically:

• -normalizecalls: normalizes function calls to be top-level only and such that arguments are identifiers (e.g. id1 = id2(id3,id4))
• -normalizereturns: normalizes return expressions to be identifiers (e.g. return id)
• -normalizepointers: normalizes pointer operations to primitive statements (id = alloc P where P is null or an integer constant, id1 = &id2, id1 = id2, id1 = *id2, *id1 = id2, orid = null)

If one or more of these options are enabled, the normalized program is printed to e.g. out/example.tip__normalized.tip.

This implementation takes advantage of many cool Scala language features that allow the code to be concise and flexible. Many of these language features are quite intuitive and easy to understand for anyone familiar with object oriented and functional programming, even without prior knowledge of Scala. Still, the following language features deserve some extra attention:

• traits
• case classes
• companion objects
• abstract type members (see e.g. GenericLattices.scala)
• implicit parameters (see e.g. TypeAnalysis.scala)
• implicit conversions (see e.g. TipType.ast2typevar)
• implicit classes (see e.g. AstNodeData.AstNodeWithDeclaration)

Tutorials and extensive documentation for Scala are available at http://docs.scala-lang.org/.

Useful tips:

• You can see what type Scala has inferred for an expression by selecting the expression and pressing Alt+Equals (depending on keyboard settings in Settings -> Keymap -> Scala -> Type Info).
• You can see what implicit conversion Scala is applying and which ones are available by selecting an expression and pressing Ctrl+Q (depending on keyboard settings in Setting -> Keymap -> Scala -> Implicit Conversion). Expressions on which an implicit is applied are underlined.

To avoid using inconsistent code styles and meaningless diffs caused by IDE reformatting we use scalafmt.

The code is automatically formatted upon compilation by SBT. In IntelliJ you need to ensure that SBT is used for compiling: in File -> Settings... -> Build Tools -> SBT, check the option "Use SBT shell for build and import".

Before committing, please double-check that all the code is in the right format by executing sbt scalafmt. By installing the scalafmt IntelliJ plugin and enabling the "Format on save" option as explained here the code is automatically formatted when the file is saved (unfortunately the formatting is only triggered when the file is explicitly saved with Ctrl-S).

• Gianluca Mezzetti
• Anders Møller

with contributions from

• Coen De Roover
• Quentin Stievenart
• Erik Krogh Kristensen
• Christian Budde Christensen