signatr-experiment

This is a skeleton project for running signatr experiments.

Installation

git clone ssh://[email protected]/PRL-PRG/signatr-experiment
cd signatr-experiment

Important: all of the following commands should be run inside the cloned repository!

• Install missing native dependencies (if applicable)

  sudo apt-get install libharfbuzz-dev libfribidi-dev

• Install R-4.0.2

  cd R-4.0.2
  curl https://cran.r-project.org/src/base/R-4/R-4.0.2.tar.gz | tar --strip-components=1 -xzf -
  ./build.sh

• Install R-dyntrace

  git clone -b r-4.0.2-signatr [email protected]:fikovnik/R-dyntrace
  cd R-dyntrace
  ./build

• Install R-dyntrace

  git clone -b r-4.0.2-signatr [email protected]:fikovnik/R-dyntrace R-dyntrace-dbg
  cd R-dyntrace-dbg
  ../build-R-dyntrace-dbg

• Clone runr

  git clone [email protected]:PRL-PRG/runr
  

• Install dependencies

  source ./env-4.0.2.sh
  ./runr/inst/install-cran-packages.R dependencies.txt

• Install runr

  make -C runr install

• Install tastr

  git clone [email protected]:PRL-PRG/tastr
  make -C tastr build

  Note that tastr needs bison-3.5.4.

• Install injectr

  git clone [email protected]:PRL-PRG/injectr
  make -C injectr install
  

• Install contractr

  git clone -b devel [email protected]:PRL-PRG/contractr
  make -C contractr install
  

• Install sxpdb

  git clone [email protected]:PRL-PRG/sxpdb
  make -C sxpdb install
  

• Install generatr

  git clone -b devel [email protected]:reallyTG/generatr.git
  make -C generatr install

• Install argtracer

  source env-dyntrace.sh
  git clone [email protected]:PRL-PRG/argtracer
  make -C argtracer install

Fuzzing

Build the DB index

> library(sxpdb)
> db <- open_db("/mnt/ocfs_vol_00/cran_db-3/", mode=TRUE)
> build_indexes(db)
NULL
> close_db(db)

Set up the environment:

source env-4.0.2.sh

The pipeline is in pipeline-fuzzing/.

Generating database

Set up the right environment:

source env-dyntrace.sh

The pipeline is in pipeline-dbgen/. To run it, start R and

targets::tar_make()

For using more cores (like, 10), do:

tar_make_future(workers = 10)

The resulting merged database will be in data/sxpdb/cran_db.

Other by-products of the pipeline:

• data/extracted-code contains the extracted files (from tests, examples and vignettes) to be run to generate the values
• data/sxpdb contains also the individual databases, one for each source file that was run