Merge pull request #1 from egraphs-good/update-readme-docker

Add Dockerfile and update README

Author: kirstenmg

Dockerfile

@@ -0,0 +1,19 @@
+# We need old ubuntu so that the package manager includes Java 6.
+FROM ubuntu:14.04.2
+
+# Install dependencies: Java 6, build essential (to build glpk), wget (to download jd-cli)
+RUN apt-get update && \
+    apt-get -y install openjdk-6-jdk build-essential wget
+
+# Download the repository
+# FIXME: we need to make the repo public for this to work
+# RUN git clone [email protected]:egraphs-good/peggy-comparison.git
+
+# # Install GLPK. We just need the glpsol executable.
+# RUN wget "https://ftp.gnu.org/gnu/glpk/glpk-5.0.tar.gz" && \
+#     unzip glpk-5.0.tar.gz && cd glpk-5.0 && \
+#     ./configure && make install && \
+#     mv examples/glpsol /usr/bin/glpsol && cd glpk-5.0 && rm -rf glpk-5.0
+
+# # Install jd-cli for command line decompiling. We need an old version to work with Java 6.
+# RUN wget "https://github.com/intoolswetrust/jd-cli/releases/download/jd-cmd-0.9.2.Final/jd-cli-0.9.2-dist.tar.gz" && tar -xzvf jd-cli-0.9.2-dist.tar.gz 

README.md

@@ -1,111 +1,18 @@
-This is the documentation for the Peggy equality saturation engine and
-compiler. Below you will find some answers to simple questions about
-how to install and use Peggy.
+This repository runs the [peggy](https://goto.ucsd.edu/~mstepp/peggy/) equality saturation engine and compiler on several example files. Its purpose is to compare peggy's optimizations to [eggcc](https://github.com/egraphs-good/eggcc), another optimizing compiler using equality saturation, on similar example files.
 
-# 0. Install Java 6.
+For the original peggy project, please refer to the [peggy documentation](https://goto.ucsd.edu/~mstepp/peggy/). This is not the original peggy project.
 
 
-## Windows
-- Install from [Oracle Java archives](https://www.oracle.com/java/technologies/javase-java-archive-javase6-downloads.html)
-(requires signup).
+# Dockerfile Setup
+The Dockerfile will clone this repository and install several dependencies.
 
-## Linux
-?
+First, build the image:
 
-## Mac
-Use a container.
 ```
-docker pull ubuntu:14.04.2
+docker build -t peggy .
 ```
 
-Create a container:
+Now run the image.
 ```
-docker run -d -v /local/path/to/peggy_1.0:/container/path/to/peggy_1.0 -w /container/path/to/peggy_1.0 --name peggy -i -t ubuntu:14.04.2 /bin/bash
+docker run -it peggy
 ```
-
-Within the container:
-```
-apt-get install openjdk-6-jdk
-```
-
-# 1. Installation.
-
-No installation is necessary. If you have the peggy_1.0.jar file, then
-you have a working copy of Peggy and all of the libraries it depends
-upon (from the lib/ folder).
-
-
-# 2. Compilation.
-
-If you wish to compile Peggy from source, you will find all the source
-code in the src/ folder. We use the Eclipse 3.1+ IDE to build Peggy.
-In Eclipse, you will need to create a new Java Project, and then add
-the code in the src/ folder to the project. You can do this either by
-copying it into the src/ folder that Eclipse creates for your new
-project, or by telling it to add the existing src/ folder to your
-project's source path. Once you have built a project and added the
-Peggy source code, you must add all the jars in the lib/ folder to the
-project's build path. Then you can refresh the project and it should
-compile correctly.
-
-
-# 3. Setup.
-
-The only setup required for Peggy is specifying the paths to a few
-external tools that it relies on. The optimizers need to know the path
-to the Pueblo, Minisat or GLPK solver program, depending on which you
-want to use. You can specify this on the command line with the
-'-minisatPath', '-puebloPath', or '-glpkPath' options.  If no path is
-specified explicitly, the path is assumed to be
-$COLLIDER_ROOT/scripts/minisat/Minisat (for minisat),
-$COLLIDER_ROOT/scripts/pueblo/Pueblo (for Pueblo), or
-/usr/bin/glpsol (for GLPK). If not using minisat, you must specify which
-solver to use by passing `minisat`, `pueblo`, or `glpk` to the `-pb` option
-on the command line.
-
-The optimizers also generate some temporary files, and you may specify
-which folder these files are created in with the option 
-'-tmpFolder <folder>'. The default is /tmp/.
-
-
-# 4. Usage.
-
-We have provided bash scripts to run the appropriate command-line
-class for each optimizer/translation validator for Java/LLVM. For any
-one of these scripts, you may call it without parameters to see a
-description of the command-line parameters it accepts.
-
-
-# 5. Quick start: Optimizing.
-
-Here's a basic example of how to optimize a Java class with Peggy.
-First, make a basic java file `Foo.java` and compile it using
-`javac Foo.java` to get the class `Foo.class`.
-Suppose we want to optimize a class "Foo.class", which is in the
-current folder.
-
-Step 1) Pick some axioms to use. There are several axiom files
-provided in the axioms/ folder, or you can write your own. To use a
-given axiom file in Peggy, simply pass the "-axioms myfile.xml"
-argument.  To provide multiple files, you may either have multiple
-"-axioms myfile.xml" pairs, or you may provide several files in one go
-with "-axioms myfile1.xml:myfile2.xml:myfile3.xml".
-         
-Step 2) To optimize Java programs, the classes you wish to optimize
-must be on the classpath. In the opt_java.sh script, the current path
-(.) is included on the classpath. We specify that we want to optimize
-Foo by adding '-O2 Foo' as a parameter to the script. This tells the
-optimizer that we want to optimize at level 2 (full optimization).
-
-Step 3) If you wish to see what axioms are applied during saturation,
-add the '-displayAxioms' parameter.
-
-Step 4) Run the script with the parameters we have determined above.
-The optimized classes will be placed in the newly-created 'optimized/'
-folder. (you may change the output folder with the '-o <folder>'
-option.
-
-
-
-
-If you have any questions/comments, please direct them to [email protected].