Dynamic Slicing of WebAssembly Binaries
Bibliographic record
Abstract
This is the replication package that accompanies the paper titled: "Dynamic Slicing of WebAssembly Binaries". # Slices dataset ## Generating the dataset The dynamic slices have been generated with [P-ORBS](https://syed-islam.github.io/research/program-analysis/#observation-based-program-slicing-orbs). The steps and scripts to generate the dynamic slices are included in the `slicing-steps/` directory. These steps also describe how to generate the `stats.csv` file that is included in this dataset. The static slices have been generated with [wassail](https://github.com/acieroid/wassail). The scripts to generate the static slices are included in the current directory (`generate-static-slices.sh` which relies on `run_wassail.sh`). These steps generate the file `static-stats.csv` that is included in this dataset. The `original-size.csv` file, included in this dataset, can be generated as follows: ```sh find subjects-wasm-extract-slice -name \*.c.wat -exec c_count {} \; | grep subjects > counts.txt echo 'slice,original fn slice' > evaluation/original-sizes.csv sed -E 's|^(.*) subjects-wasm-extract-slice/[^/]*/([^/]*)/.*$|\2,\1|' counts.txt >> evaluation/original-sizes.csv ``` The numbers in Table 1 of the paper (the list of programs in the dataset along with their sizes) can be generated as follows. For the WebAssembly files, we can count the function size: ``` find subjects-wasm-extract-slice -name \*.c.wat -exec c_count {} \; | grep subjects > counts.txt sed -E 's|^(.*) subjects-wasm-extract-slice/([^/]*)/.*$|\1 \2|' counts.txt | python evaluation/table1-wasm-mean.py ``` or the full program size: ``` find subjects-wasm-extract-slice -name t.wat -exec c_count {} \; | grep subjects > counts.txt sed -E 's|^(.*) subjects-wasm-extract-slice/([^/]*)/.*$|\1 \2|' counts.txt | python evaluation/table1-wasm-mean.py ``` ## Structure of the dataset The dataset is structured as follows: - `subjects/` contains the instrumented `.c` source code, along with scripts to generate the dynamic slices. The original source code can be obtained by removing the line `printf("\nORBS:%x\n....`. - `subjects-wasm-extract-slice/` contains the original WebAssembly programs to slice. Each program has two files: `t.wat` is the full binary file, and `name.c.wat` is the binary code of the function containing the slicing criterion. - `all_slices/` contains the slices. For example, program `adpcm_ah1_254_expr` has the following files - `adpcm/adpcm_ah1_254_expr/EWS_adpcm.wat`: the EWS slice - `adpcm/adpcm_ah1_254_expr/SEW_adpcm.wat`: the SEW slice - `adpcm/adpcm_ah1_254_expr/ESW_adpcm.wat`: the ESW slice - `adpcm/adpcm_ah1_254_expr/static_adpcm.wat.slice`: the SWS slice The other files are produced by intermediary steps and can be ignored. They are: - `adpcm/adpcm_ah1_254_expr/ESW_adpcm.wat.orig`: original (unsliced) binary *file* from which SW and ESW slices are computed - `adpcm/adpcm_ah1_254_expr/SEW_adpcm.wat.orig`: original (unsliced) binary *function* from which SEW slice is computed - `adpcm/adpcm_ah1_254_expr/SW_adpcm.wat`: slice of entire binary file from which ESW slice is extracted - `adpcm/adpcm_ah1_254_expr/WS_adpcm.wat`: compiled (binary) version of dynamic C slice from which EWS slice is extracted # Research questions ## RQ1 The script `./RQ1.py` found in the `evaluation/` directory generates: - Figure 3 (time.pdf) - The mean, min, max, and stddev of the times - How many slices are computed below 10, 100, 1000, and 10000 seconds ## RQ2 The script `./RQ2.py` found in the `evaluation/` directory generates: - Figure 4 (loc.pdf) - The mean, median, min, max, and stddev of the sizes - The largest differences between the approaches - The number of slices larger than the original program ## RQ3 and RQ4 The process for these research questions is manual and requires comparing slices. It cannot be automated. We did make heavy use of `diff --side-by-side` in this analysis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.006 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".