robojerk/codeql-action

Fix out of memory in hash computation

Browse source
This commit is contained in:
Edoardo Pirovano 2021-06-07 12:07:56 +01:00 • committed by Edoardo Pirovano
parent 242fd828aa
commit 02e8dcfe9c
10 changed files with 151 additions and 91 deletions
Download patch file Download diff file Expand all files Collapse all files

34
lib/fingerprints.js generated
View file

@ -36,9 +36,9 @@ function computeFirstMod() {
* the hashes of the lines near the end of the file.
*
* @param callback function that is called with the line number (1-based) and hash for every line
* @param input The file's contents
* @param filepath The path to the file to hash
*/
function hash(callback, input) {
async function hash(callback, filepath) {
// A rolling view in to the input
const window = Array(BLOCK_SIZE).fill(0);
// If the character in the window is the start of a new line
@ -82,12 +82,11 @@ function hash(callback, input) {
// as we go. Once we reach a point in the window again then we've processed
// BLOCK_SIZE characters and if the last character at this point in the window
// was the start of a line then we should output the hash for that line.
for (let i = 0, len = input.length; i <= len; i++) {
let current = i === len ? 65535 : input.charCodeAt(i);
const processCharacter = function (current) {
// skip tabs, spaces, and line feeds that come directly after a carriage return
if (current === space || current === tab || (prevCR && current === lf)) {
prevCR = false;
continue;
return;
}
// replace CR with LF
if (current === cr) {
@ -109,7 +108,18 @@ function hash(callback, input) {
lineStart = true;
}
updateHash(current);
}
};
await new Promise((fulfill) => {
const readStream = fs.createReadStream(filepath, "utf8");
readStream.on("close", fulfill);
readStream.on("end", () => {
processCharacter(65535);
});
readStream.on("data", (data) => {
for (let i = 0; i < data.length; ++i)
processCharacter(data.charCodeAt(i));
});
});
// Flush the remaining lines
for (let i = 0; i < BLOCK_SIZE; i++) {
if (lineNumbers[index] !== -1) {
@ -206,8 +216,8 @@ function resolveUriToFile(location, artifacts, checkoutPath, logger) {
exports.resolveUriToFile = resolveUriToFile;
// Compute fingerprints for results in the given sarif file
// and return an updated sarif file contents.
function addFingerprints(sarifContents, checkoutPath, logger) {
var _a, _b;
async function addFingerprints(sarifContents, checkoutPath, logger) {
var _a, _b, _c, _d, _e;
const sarif = JSON.parse(sarifContents);
// Gather together results for the same file and construct
// callbacks to accept hashes for that file and update the location
@ -222,6 +232,11 @@ function addFingerprints(sarifContents, checkoutPath, logger) {
logger.debug(`Unable to compute fingerprint for invalid location: ${JSON.stringify(primaryLocation)}`);
continue;
}
if (typeof ((_e = (_d = (_c = primaryLocation) === null || _c === void 0 ? void 0 : _c.physicalLocation) === null || _d === void 0 ? void 0 : _d.region) === null || _e === void 0 ? void 0 : _e.startLine) ===
"undefined") {
// Locations without a line number are unlikely to be source files
continue;
}
const filepath = resolveUriToFile(primaryLocation.physicalLocation.artifactLocation, artifacts, checkoutPath, logger);
if (!filepath) {
continue;
@ -240,8 +255,7 @@ function addFingerprints(sarifContents, checkoutPath, logger) {
c(lineNumber, hashValue);
}
};
const fileContents = fs.readFileSync(filepath).toString();
hash(teeCallback, fileContents);
await hash(teeCallback, filepath);
}
return JSON.stringify(sarif);
}

2
lib/fingerprints.js.map
View file

File diff suppressed because one or more lines are too long

51
lib/fingerprints.test.js generated
View file

@ -16,28 +16,33 @@ const ava_1 = __importDefault(require("ava"));
const fingerprints = __importStar(require("./fingerprints"));
const logging_1 = require("./logging");
const testing_utils_1 = require("./testing-utils");
const util = __importStar(require("./util"));
testing_utils_1.setupTests(ava_1.default);
function testHash(t, input, expectedHashes) {
let index = 0;
const callback = function (lineNumber, hash) {
t.is(lineNumber, index + 1);
t.is(hash, expectedHashes[index]);
index++;
};
fingerprints.hash(callback, input);
t.is(index, input.split(/\r\n|\r|\n/).length);
async function testHash(t, input, expectedHashes) {
await util.withTmpDir(async (tmpDir) => {
const tmpFile = path.resolve(tmpDir, "testfile");
fs.writeFileSync(tmpFile, input);
let index = 0;
const callback = function (lineNumber, hash) {
t.is(lineNumber, index + 1);
t.is(hash, expectedHashes[index]);
index++;
};
await fingerprints.hash(callback, tmpFile);
t.is(index, input.split(/\r\n|\r|\n/).length);
});
}
ava_1.default("hash", (t) => {
ava_1.default("hash", async (t) => {
// Try empty file
testHash(t, "", ["c129715d7a2bc9a3:1"]);
await testHash(t, "", ["c129715d7a2bc9a3:1"]);
// Try various combinations of newline characters
testHash(t, " a\nb\n \t\tc\n d", [
await testHash(t, " a\nb\n \t\tc\n d", [
"271789c17abda88f:1",
"54703d4cd895b18:1",
"180aee12dab6264:1",
"a23a3dc5e078b07b:1",
]);
testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End", [
await testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End", [
"8b7cf3e952e7aeb2:1",
"b1ae1287ec4718d9:1",
"bff680108adb0fcc:1",
@ -45,7 +50,7 @@ ava_1.default("hash", (t) => {
"b86d3392aea1be30:1",
"e6ceba753e1a442:1",
]);
testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End\n", [
await testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -54,7 +59,7 @@ ava_1.default("hash", (t) => {
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\nworld!!!\r\r\r \t\tGreetings\r End\r", [
await testHash(t, " hello; \t\nworld!!!\r\r\r \t\tGreetings\r End\r", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -63,7 +68,7 @@ ava_1.default("hash", (t) => {
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\r\nworld!!!\r\n\r\n\r\n \t\tGreetings\r\n End\r\n", [
await testHash(t, " hello; \t\r\nworld!!!\r\n\r\n\r\n \t\tGreetings\r\n End\r\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -72,7 +77,7 @@ ava_1.default("hash", (t) => {
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\nworld!!!\r\n\n\r \t\tGreetings\r End\r\n", [
await testHash(t, " hello; \t\nworld!!!\r\n\n\r \t\tGreetings\r End\r\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -82,7 +87,7 @@ ava_1.default("hash", (t) => {
"c129715d7a2bc9a3:1",
]);
// Try repeating line that will generate identical hashes
testHash(t, "Lorem ipsum dolor sit amet.\n".repeat(10), [
await testHash(t, "Lorem ipsum dolor sit amet.\n".repeat(10), [
"a7f2ff13bc495cf2:1",
"a7f2ff13bc495cf2:2",
"a7f2ff13bc495cf2:3",
@ -95,7 +100,7 @@ ava_1.default("hash", (t) => {
"cc97dc7b1d7d8f7b:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, "x = 2\nx = 1\nprint(x)\nx = 3\nprint(x)\nx = 4\nprint(x)\n", [
await testHash(t, "x = 2\nx = 1\nprint(x)\nx = 3\nprint(x)\nx = 4\nprint(x)\n", [
"e54938cc54b302f1:1",
"bb609acbe9138d60:1",
"1131fd5871777f34:1",
@ -150,7 +155,7 @@ ava_1.default("resolveUriToFile", (t) => {
t.is(testResolveUriToFile(dirpath, undefined, []), undefined);
t.is(testResolveUriToFile(`file://${dirpath}`, undefined, []), undefined);
});
ava_1.default("addFingerprints", (t) => {
ava_1.default("addFingerprints", async (t) => {
// Run an end-to-end test on a test file
let input = fs
.readFileSync(`${__dirname}/../src/testdata/fingerprinting.input.sarif`)
@ -163,9 +168,9 @@ ava_1.default("addFingerprints", (t) => {
expected = JSON.stringify(JSON.parse(expected));
// The URIs in the SARIF files resolve to files in the testdata directory
const checkoutPath = path.normalize(`${__dirname}/../src/testdata`);
t.deepEqual(fingerprints.addFingerprints(input, checkoutPath, logging_1.getRunnerLogger(true)), expected);
t.deepEqual(await fingerprints.addFingerprints(input, checkoutPath, logging_1.getRunnerLogger(true)), expected);
});
ava_1.default("missingRegions", (t) => {
ava_1.default("missingRegions", async (t) => {
// Run an end-to-end test on a test file
let input = fs
.readFileSync(`${__dirname}/../src/testdata/fingerprinting2.input.sarif`)
@ -178,6 +183,6 @@ ava_1.default("missingRegions", (t) => {
expected = JSON.stringify(JSON.parse(expected));
// The URIs in the SARIF files resolve to files in the testdata directory
const checkoutPath = path.normalize(`${__dirname}/../src/testdata`);
t.deepEqual(fingerprints.addFingerprints(input, checkoutPath, logging_1.getRunnerLogger(true)), expected);
t.deepEqual(await fingerprints.addFingerprints(input, checkoutPath, logging_1.getRunnerLogger(true)), expected);
});
//# sourceMappingURL=fingerprints.test.js.map

2
lib/fingerprints.test.js.map
View file

File diff suppressed because one or more lines are too long

2
lib/upload-lib.js generated
View file

@ -243,7 +243,7 @@ async function uploadFiles(sarifFiles, repositoryNwo, commitOid, ref, analysisKe
validateSarifFileSchema(file, logger);
}
let sarifPayload = combineSarifFiles(sarifFiles);
sarifPayload = fingerprints.addFingerprints(sarifPayload, checkoutPath, logger);
sarifPayload = await fingerprints.addFingerprints(sarifPayload, checkoutPath, logger);
sarifPayload = populateRunAutomationDetails(sarifPayload, category, analysisKey, environment);
const zippedSarif = zlib_1.default.gzipSync(sarifPayload).toString("base64");
const checkoutURI = file_url_1.default(checkoutPath);

2
lib/upload-lib.js.map
View file

File diff suppressed because one or more lines are too long

105
src/fingerprints.test.ts
View file

@ -7,32 +7,41 @@ import test from "ava";
import * as fingerprints from "./fingerprints";
import { getRunnerLogger } from "./logging";
import { setupTests } from "./testing-utils";
import * as util from "./util";
setupTests(test);
function testHash(t: ava.Assertions, input: string, expectedHashes: string[]) {
let index = 0;
const callback = function (lineNumber: number, hash: string) {
t.is(lineNumber, index + 1);
t.is(hash, expectedHashes[index]);
index++;
};
fingerprints.hash(callback, input);
t.is(index, input.split(/\r\n|\r|\n/).length);
async function testHash(
t: ava.Assertions,
input: string,
expectedHashes: string[]
) {
await util.withTmpDir(async (tmpDir) => {
const tmpFile = path.resolve(tmpDir, "testfile");
fs.writeFileSync(tmpFile, input);
let index = 0;
const callback = function (lineNumber: number, hash: string) {
t.is(lineNumber, index + 1);
t.is(hash, expectedHashes[index]);
index++;
};
await fingerprints.hash(callback, tmpFile);
t.is(index, input.split(/\r\n|\r|\n/).length);
});
}
test("hash", (t: ava.Assertions) => {
test("hash", async (t: ava.Assertions) => {
// Try empty file
testHash(t, "", ["c129715d7a2bc9a3:1"]);
await testHash(t, "", ["c129715d7a2bc9a3:1"]);
// Try various combinations of newline characters
testHash(t, " a\nb\n \t\tc\n d", [
await testHash(t, " a\nb\n \t\tc\n d", [
"271789c17abda88f:1",
"54703d4cd895b18:1",
"180aee12dab6264:1",
"a23a3dc5e078b07b:1",
]);
testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End", [
await testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End", [
"8b7cf3e952e7aeb2:1",
"b1ae1287ec4718d9:1",
"bff680108adb0fcc:1",
@ -40,7 +49,7 @@ test("hash", (t: ava.Assertions) => {
"b86d3392aea1be30:1",
"e6ceba753e1a442:1",
]);
testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End\n", [
await testHash(t, " hello; \t\nworld!!!\n\n\n \t\tGreetings\n End\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -49,7 +58,7 @@ test("hash", (t: ava.Assertions) => {
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\nworld!!!\r\r\r \t\tGreetings\r End\r", [
await testHash(t, " hello; \t\nworld!!!\r\r\r \t\tGreetings\r End\r", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -58,16 +67,20 @@ test("hash", (t: ava.Assertions) => {
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\r\nworld!!!\r\n\r\n\r\n \t\tGreetings\r\n End\r\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
"e20e36e16fcb0cc8:1",
"b3edc88f2938467e:1",
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]);
testHash(t, " hello; \t\nworld!!!\r\n\n\r \t\tGreetings\r End\r\n", [
await testHash(
t,
" hello; \t\r\nworld!!!\r\n\r\n\r\n \t\tGreetings\r\n End\r\n",
[
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
"e20e36e16fcb0cc8:1",
"b3edc88f2938467e:1",
"c8e28b0b4002a3a0:1",
"c129715d7a2bc9a3:1",
]
);
await testHash(t, " hello; \t\nworld!!!\r\n\n\r \t\tGreetings\r End\r\n", [
"e9496ae3ebfced30:1",
"fb7c023a8b9ccb3f:1",
"ce8ba1a563dcdaca:1",
@ -78,7 +91,7 @@ test("hash", (t: ava.Assertions) => {
]);
// Try repeating line that will generate identical hashes
testHash(t, "Lorem ipsum dolor sit amet.\n".repeat(10), [
await testHash(t, "Lorem ipsum dolor sit amet.\n".repeat(10), [
"a7f2ff13bc495cf2:1",
"a7f2ff13bc495cf2:2",
"a7f2ff13bc495cf2:3",
@ -92,16 +105,20 @@ test("hash", (t: ava.Assertions) => {
"c129715d7a2bc9a3:1",
]);
testHash(t, "x = 2\nx = 1\nprint(x)\nx = 3\nprint(x)\nx = 4\nprint(x)\n", [
"e54938cc54b302f1:1",
"bb609acbe9138d60:1",
"1131fd5871777f34:1",
"5c482a0f8b35ea28:1",
"54517377da7028d2:1",
"2c644846cb18d53e:1",
"f1b89f20de0d133:1",
"c129715d7a2bc9a3:1",
]);
await testHash(
t,
"x = 2\nx = 1\nprint(x)\nx = 3\nprint(x)\nx = 4\nprint(x)\n",
[
"e54938cc54b302f1:1",
"bb609acbe9138d60:1",
"1131fd5871777f34:1",
"5c482a0f8b35ea28:1",
"54517377da7028d2:1",
"2c644846cb18d53e:1",
"f1b89f20de0d133:1",
"c129715d7a2bc9a3:1",
]
);
});
function testResolveUriToFile(uri: any, index: any, artifactsURIs: any[]) {
@ -170,7 +187,7 @@ test("resolveUriToFile", (t) => {
t.is(testResolveUriToFile(`file://${dirpath}`, undefined, []), undefined);
});
test("addFingerprints", (t) => {
test("addFingerprints", async (t) => {
// Run an end-to-end test on a test file
let input = fs
.readFileSync(`${__dirname}/../src/testdata/fingerprinting.input.sarif`)
@ -187,12 +204,16 @@ test("addFingerprints", (t) => {
const checkoutPath = path.normalize(`${__dirname}/../src/testdata`);
t.deepEqual(
fingerprints.addFingerprints(input, checkoutPath, getRunnerLogger(true)),
await fingerprints.addFingerprints(
input,
checkoutPath,
getRunnerLogger(true)
),
expected
);
});
test("missingRegions", (t) => {
test("missingRegions", async (t) => {
// Run an end-to-end test on a test file
let input = fs
.readFileSync(`${__dirname}/../src/testdata/fingerprinting2.input.sarif`)
@ -209,7 +230,11 @@ test("missingRegions", (t) => {
const checkoutPath = path.normalize(`${__dirname}/../src/testdata`);
t.deepEqual(
fingerprints.addFingerprints(input, checkoutPath, getRunnerLogger(true)),
await fingerprints.addFingerprints(
input,
checkoutPath,
getRunnerLogger(true)
),
expected
);
});

38
src/fingerprints.ts
View file

@ -34,9 +34,9 @@ type hashCallback = (lineNumber: number, hash: string) => void;
* the hashes of the lines near the end of the file.
*
* @param callback function that is called with the line number (1-based) and hash for every line
* @param input The file's contents
* @param filepath The path to the file to hash
*/
export function hash(callback: hashCallback, input: string) {
export async function hash(callback: hashCallback, filepath: string) {
// A rolling view in to the input
const window = Array(BLOCK_SIZE).fill(0);
@ -87,12 +87,11 @@ export function hash(callback: hashCallback, input: string) {
// as we go. Once we reach a point in the window again then we've processed
// BLOCK_SIZE characters and if the last character at this point in the window
// was the start of a line then we should output the hash for that line.
for (let i = 0, len = input.length; i <= len; i++) {
let current = i === len ? 65535 : input.charCodeAt(i);
const processCharacter = function (current: number) {
// skip tabs, spaces, and line feeds that come directly after a carriage return
if (current === space || current === tab || (prevCR && current === lf)) {
prevCR = false;
continue;
return;
}
// replace CR with LF
if (current === cr) {
@ -113,7 +112,19 @@ export function hash(callback: hashCallback, input: string) {
lineStart = true;
}
updateHash(current);
}
};
await new Promise((fulfill) => {
const readStream = fs.createReadStream(filepath, "utf8");
readStream.on("close", fulfill);
readStream.on("end", () => {
processCharacter(65535);
});
readStream.on("data", (data) => {
for (let i = 0; i < data.length; ++i)
processCharacter(data.charCodeAt(i));
});
});
// Flush the remaining lines
for (let i = 0; i < BLOCK_SIZE; i++) {
@ -237,11 +248,11 @@ export function resolveUriToFile(
// Compute fingerprints for results in the given sarif file
// and return an updated sarif file contents.
export function addFingerprints(
export async function addFingerprints(
sarifContents: string,
checkoutPath: string,
logger: Logger
): string {
): Promise<string> {
const sarif = JSON.parse(sarifContents);
// Gather together results for the same file and construct
@ -263,6 +274,14 @@ export function addFingerprints(
continue;
}
if (
typeof primaryLocation?.physicalLocation?.region?.startLine ===
"undefined"
) {
// Locations without a line number are unlikely to be source files
continue;
}
const filepath = resolveUriToFile(
primaryLocation.physicalLocation.artifactLocation,
artifacts,
@ -289,8 +308,7 @@ export function addFingerprints(
c(lineNumber, hashValue);
}
};
const fileContents = fs.readFileSync(filepath).toString();
hash(teeCallback, fileContents);
await hash(teeCallback, filepath);
}
return JSON.stringify(sarif);

4
src/testdata/fingerprinting2.expected.sarif vendored
View file

@ -30,9 +30,7 @@
"message": {
"text": "This header file should contain a header guard to prevent multiple inclusion."
},
"partialFingerprints": {
"primaryLocationLineHash": "599c824c91d0f75e:1"
},
"partialFingerprints": {},
"ruleId": "cpp/missing-header-guard",
"ruleIndex": 0
}

2
src/upload-lib.ts
View file

@ -356,7 +356,7 @@ async function uploadFiles(
}
let sarifPayload = combineSarifFiles(sarifFiles);
sarifPayload = fingerprints.addFingerprints(
sarifPayload = await fingerprints.addFingerprints(
sarifPayload,
checkoutPath,
logger