Move data files into data/, separate transcripts from PDFs in docs/ #4

Workflow file for this run

	# CI workflow for the CSCI 331 zipcode file-structures project.
	#
	# There is no Makefile or CMakeLists.txt in this repository (see README.md,
	# "Environment Setup"). The build commands below are copied verbatim from
	# that section, which itself was reverse-engineered from the project's own
	# committed script(1) terminal transcripts (docs/transcripts/req3_script.txt,
	# docs/transcripts/req4_script.txt, docs/transcripts/req7_script.txt,
	# docs/transcripts/req8_script.txt). Do not "simplify" these into a single
	# g++ invocation or a Makefile here - the point of this workflow is to keep
	# building the project exactly the way the README documents it.

	name: CI

	# Run on every push, to any branch, and allow manually triggering a run
	# from the Actions tab.
	on:
	push:
	workflow_dispatch:

	jobs:
	build:
	# ubuntu-latest ships g++ preinstalled, which is all this project needs
	# (no third-party libraries, just the C++ standard library).
	runs-on: ubuntu-latest

	steps:
	# Check out the repository so the compiler has the source files and
	# the committed data/ .csv files to work with.
	- name: Check out repository
	uses: actions/checkout@v4

	- name: Show g++ version
	run: g++ --version

	# --- Build target 1: proj3_original ---
	# Project 3, original version: reads a plain .csv file directly
	# through the simple windowed Buffer/Zipcode classes.
	- name: Compile proj3_original (Project 3, CSV-based, no header/index)
	run: g++ -o proj3_original src/proj3.cpp src/buffer.cpp src/zipcode.cpp

	# --- Build target 2: proj3_updated ---
	# Project 3 logic re-targeted at the .zcr format (Buffer2/Zipcode2).
	- name: Compile proj3_updated (Project 3 logic against the .zcr format)
	run: g++ -o proj3_updated src/proj3Updated.cpp src/buffer2.cpp src/zipcode2.cpp

	# --- Build target 3: converter ---
	# Converts a .csv file into the project's custom .zcr (header +
	# length-indicated records) format. This is the program containing
	# the fileConverter.cpp fix for the uncaught std::out_of_range
	# exception on filenames without a "." (see README, Security /
	# Robustness Findings).
	- name: Compile converter (CSV to ZCR file converter)
	run: g++ -o converter src/fileConverter.cpp src/headerRecord.cpp src/headerBuffer.cpp src/buffer.cpp src/zipcode.cpp

	# --- Build target 4: proj4_norm (representative of proj4_norm / proj4_norm_place / proj4_rand / proj4_rand_state) ---
	# The README explicitly documents that all four proj4_* binaries are
	# the SAME source combination, compiled four separate times, with the
	# only difference being which renamed data-file variant
	# (data/us_postal_codes.zcr/.idx) happens to be present on disk at
	# build time. Since the source and compile command are identical for all
	# four, compiling it once here fully verifies that source combination;
	# repeating the identical g++ command three more times under
	# different -o names would not exercise any additional code. This is
	# also the binary containing the headerRecord.cpp fix for the
	# out-of-bounds vector write in HeaderRecord::Unpack().
	- name: Compile proj4_norm (findZipcode + primaryKeyIndex against .zcr/.idx)
	run: g++ -o proj4_norm src/findZipcode.cpp src/primaryKeyIndex.cpp src/headerRecord.cpp src/headerBuffer.cpp src/buffer2.cpp src/zipcode2.cpp

	# --- Runtime smoke test for the fileConverter.cpp fix ---
	# fileConverter.cpp prompts interactively ("Enter the name of the CSV
	# file: ") and reads the filename with `std::cin >> csvFileName`
	# rather than taking it as a command-line argument. That prompt/read
	# is still a simple, single-value, non-interactive-scriptable input:
	# piping one line into stdin reproduces the exact documented usage
	# without needing any multi-step interactive session.
	#
	# Case A: a filename with no "." at all. This is the exact input
	# (`sfsdfsdfsdfsdf`) recorded in the project's own transcript
	# (docs/transcripts/req4_script.txt:153-158) that used to crash the program with
	# `terminate called after throwing an instance of 'std::out_of_range'`
	# (SIGABRT). After the fix, csvFileName.rfind(".") returning npos is
	# checked before substr() is called, so the program should instead
	# print the documented "not a CSV file" error and exit cleanly with
	# status 1. We assert the process does not abort (non-zero exit code
	# from a signal, e.g. 134 for SIGABRT) and instead exits with the
	# expected status 1.
	- name: Runtime smoke test - converter with a filename lacking "."
	run: \|
	set +e
	output=$(echo "sfsdfsdfsdfsdf" \| ./converter 2>&1)
	status=$?
	echo "$output"
	if [ $status -eq 134 ]; then
	echo "FAIL: converter aborted (SIGABRT) on a filename with no '.' - the fileConverter.cpp fix appears to be missing."
	exit 1
	fi
	if [ $status -ne 1 ]; then
	echo "FAIL: expected converter to exit with status 1 (graceful rejection), got $status."
	exit 1
	fi
	if ! echo "$output" \| grep -q "Input file is not a CSV file"; then
	echo "FAIL: expected the documented 'Input file is not a CSV file.' message."
	exit 1
	fi
	echo "PASS: converter rejected a filename with no '.' cleanly instead of crashing."

	# Case B: a normal, well-formed conversion using one of the committed
	# small sample CSV files (data/test1.csv), matching the README's
	# documented workflow of running converter against a .csv file to
	# produce a .zcr file. This exercises the header-writing code path in
	# headerRecord.cpp/headerBuffer.cpp on a successful run, in addition
	# to the error path exercised above.
	- name: Runtime smoke test - converter with a valid CSV file
	run: \|
	set +e
	output=$(echo "data/test1.csv" \| ./converter 2>&1)
	status=$?
	echo "$output"
	if [ $status -ne 0 ]; then
	echo "FAIL: expected converter to exit 0 on a valid CSV file, got $status."
	exit 1
	fi
	if ! echo "$output" \| grep -q "File conversion complete"; then
	echo "FAIL: expected the documented 'File conversion complete...' message."
	exit 1
	fi
	echo "PASS: converter completed a normal CSV to ZCR conversion."

	# Note on headerRecord.cpp's out-of-bounds vector write fix: that bug
	# only manifests when a .zcr/.idx header is corrupted or crafted to
	# claim more fields (fieldsPerRecord) than HeaderRecord was
	# constructed with. Every header produced by this codebase's own
	# tools always matches (fieldsPerRecord == 6), so reproducing the bug
	# requires hand-crafting a malformed header file - a multi-step setup
	# rather than a single documented command-line invocation. Per the
	# task scope, that scenario is left to compile-time verification only
	# (the proj4_norm and converter builds above both compile
	# headerRecord.cpp) rather than a scripted runtime repro here.

Navigation Menu

Move data files into data/, separate transcripts from PDFs in docs/ #4

Workflow file

Move data files into data/, separate transcripts from PDFs in docs/ #4

Uh oh!

Workflow file for this run