fs.read into zero buffer should not throw exception #4518

feross · 2016-01-03T16:28:57Z

feross · 2016-01-03T16:33:53Z

Sign in to view

@@ -1098,7 +1098,7 @@ static void Read(const FunctionCallbackInfo<Value>& args) {
  size_t buffer_length = Buffer::Length(buffer_obj);

  size_t off = args[2]->Int32Value();
-  if (off >= buffer_length) {
+  if (off > buffer_length) {


This should be safe because the following if-statement will catch if offset + length would be out-of-bounds of the buffer.

I don't think it is. If buffer_length == 0, then it's possible for buffer_data == nullptr. nullptr + 0 is technically defined behavior (in C++, not C) but passing that pointer onto the kernel can - but does not have to - fail.

If zero-sized reads are desirable, I would handle them in lib/fs.js.

Trott · 2016-01-03T16:51:20Z

CI: https://ci.nodejs.org/job/node-test-commit/1611/

thefourtheye · 2016-01-03T17:09:03Z

Sign in to view

+    bufferSync = new Buffer(0),
+    readCalled = 0;
+
+fs.read(fd, bufferAsync, 0, 0, 0, function(err, bytesRead) {


Use common.mustCall instead of using readCalled

I just copied this from another test. Will update.

thefourtheye · 2016-01-03T17:12:55Z

What could be the possible use-case which would need this?

Trott · 2016-01-03T17:17:45Z

@thefourtheye webtorrent/webtorrent#461

feross · 2016-01-03T17:19:49Z

@thefourtheye Yep, what @Trott said.

Zero-length files are valid files. If the user passes one into our module, I think it's reasonable to expect fs.read to work. Instead, we currently have to resort to detecting this special case like this: random-access-storage/random-access-file@32f3d1b

feross · 2016-01-03T17:25:37Z

@thefourtheye Just addressed your comment by updating the PR.

tejasmanohar · 2016-01-03T18:11:28Z

👍

Trott · 2016-01-03T22:17:38Z

Re-running CI since force push: https://ci.nodejs.org/job/node-test-commit/1612/

feross · 2016-01-03T23:01:33Z

I just force pushed again to address @bnoordhuis's comments. I moved the logic to lib/fs.js, as he suggested.

@Trott You might want to re-run the CI again ;)

Trott · 2016-01-03T23:04:39Z

CI round three: https://ci.nodejs.org/job/node-test-pull-request/1136/

evanlucas · 2016-01-04T12:02:23Z

CI looks good (minus the occasional flaky tests). LGTM

thefourtheye · 2016-01-04T12:13:39Z

lib/fs.js

@@ -609,6 +609,12 @@ fs.read = function(fd, buffer, offset, length, position, callback) {
    };
  }

+  if (length === 0) {
+    return process.nextTick(function() {


Hmmm, this is not right. There is actually a deprecated fs.read form which returns string.

So,

This will actually return a buffer and the deprecated fs.read will collapse.

The order of parameters to the callback, in string form, is actually string and the bytesRead.

Actually, this calls the redefined callback function (that starts on line 600), so the right arguments should be passed to the callback function, namely: null, '', and 0.

Unless I'm missing something.

You comment prompted me to check the fs.readSync case and the issue you describe exists there. I'll fix it and add a test.

feross · 2016-01-04T15:44:02Z

Caught one additional bug in the legacy usage of fs.readSync, so I fixed it and added a test.

Please take a look at the changes and re-run CI. 👍

thefourtheye · 2016-01-04T17:17:23Z

Sign in to view

+    fd = fs.openSync(filepath, 'r'),
+    expected = '';
+
+fs.read(fd, 0, 0, 'utf-8', common.mustCall(function(err, str, bytesRead) {


Make length consistent with this and the readSync case.

thefourtheye · 2016-01-04T17:21:15Z

lib/fs.js

@@ -648,6 +654,14 @@ fs.readSync = function(fd, buffer, offset, length, position) {
    offset = 0;
  }

+  if (length === 0) {


The check here means that even if the file doesn't exist/inaccessible, if the length is zero, we will return valid values.

How do you propose to solve this? @bnoordhuis asked me to move the checks to lib/fs.js to avoid some kernel edge case.

Thoughts?

@thefourtheye Since the fd itself is passed best we can attempt to do is read directly from it. But as @feross mentioned, there are cross platform issues that prevents this being done correctly. I don't think that assuming the fd is alright is a problem in this case.

@trevnorris Ah, you are right. I totally missed that. Then this will not be a problem here.

jasnell · 2016-01-05T18:46:58Z

Defensively marking as semver-major due to the error handling change. Minor might be more appropriate.

feross · 2016-01-06T14:36:54Z

Updated PR to make the test more consistent, per @thefourtheye's comment.

I think this is ready to be landed 👍

thefourtheye · 2016-01-06T14:41:09Z

CI Run: https://ci.nodejs.org/job/node-test-pull-request/1149/

jasnell · 2016-01-06T15:52:04Z

Sign in to view

+'use strict';
+var common = require('../common');
+var assert = require('assert');
+var path = require('path'),


these should use const

You guys should really add an eslint rule for this. So much overhead for contributors to deal with the back-and-forth.

As far as I know, the eslint rule only catches const vs let, not vs var..

cc @silverwind?

jasnell · 2016-01-06T15:52:37Z

Sign in to view

+    fd = fs.openSync(filepath, 'r'),
+    bufferAsync = new Buffer(0),
+    bufferSync = new Buffer(0);
+


Nit: would prefer the decl's to be separated but not critical

This was copied from another test file. Will change.

jasnell · 2016-01-06T15:53:10Z

Sign in to view

+var common = require('../common');
+var assert = require('assert');
+var path = require('path'),
+    fs = require('fs'),


const please :-)

Fixes #4517.

feross · 2016-01-06T16:41:54Z

Force pushed again to address @jasnell's comments.

fingers crossed

jasnell · 2016-01-06T16:50:57Z

thank you @feross ... and thank you for bearing with us ;-) the contribution is very appreciated.
Doing another CI pass: https://ci.nodejs.org/job/node-test-pull-request/1151/
LGTM if CI is green! :-)

feross · 2016-01-06T16:55:27Z

Thanks @jasnell, I appreciate the attention to detail.

jasnell · 2016-01-06T17:28:47Z

CI failures look unrelated! Will get this landed

Fixes: #4517. PR-URL: #4518 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com>

jasnell · 2016-01-06T17:34:50Z

Landed in 2b15e68

feross · 2016-01-06T17:35:40Z

Nice!

Fixes: nodejs#4517. PR-URL: nodejs#4518 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com>

The following significant (semver-major) changes have been made since the previous Node v5.0.0 release. * Buffer * New Buffer constructors have been added [#4682](#4682) * Previously deprecated Buffer APIs are removed [#5048](#5048), [#4594](#4594) * Improved error handling [#4514](#4514) * Cluster * Worker emitted as first argument in 'message' event [#5361](#5361). * Crypto * Improved error handling [#3100](#3100), [#5611](#5611) * Simplified Certificate class bindings [#5382](#5382) * Improved control over FIPS mode [#5181](#5181) * pbkdf2 digest overloading is deprecated [#4047](#4047) * Dependencies * Reintroduce shared c-ares build support [#5775](#5775). * V8 updated to 5.0.71.31 [#6111](#6111). * DNS * Add resolvePtr API to query plain DNS PTR records [#4921](#4921). * Domains * Clear stack when no error handler [#4659](#4659). * File System * The `fs.realpath()` and `fs.realpathSync()` methods have been updated to use a more efficient libuv implementation. This change includes the removal of the `cache` argument and the method can throw new errors [#3594](#3594) * FS apis can now accept and return paths as Buffers [#5616](#5616). * Error handling and type checking improvements [#5616](#5616), [#5590](#5590), [#4518](#4518), [#3917](#3917). * fs.read's string interface is deprecated [#4525](#4525) * HTTP * 'clientError' can now be used to return custom errors from an HTTP server [#4557](#4557). * Modules * Current directory is now prioritized for local lookups [#5689](#5689) * Symbolic links are preserved when requiring modules [#5950](#5950) * Net * DNS hints no longer implicitly set [#6021](#6021). * Improved error handling and type checking [#5981](#5981), [#5733](#5733), [#2904](#2904) * Path * Improved type checking [#5348](#5348). * Process * Introduce process warnings API [#4782](#4782). * Throw exception when non-function passed to nextTick [#3860](#3860). * Readline * Emit key info unconditionally [#6024](#6024) * REPL * Assignment to `_` will emit a warning. [#5535](#5535) * Timers * Fail early when callback is not a function [#4362](#4362) * TLS * Rename 'clientError' to 'tlsClientError' [#4557](#4557) * SHA1 used for sessionIdContext [#3866](#3866) * TTY * Previously deprecated setRawMode wrapper is removed [#2528](#2528). * Util * Changes to Error object formatting [#4582](#4582). * Windows * Windows XP and Vista are no longer supported [#5167](#5167), [#5167](#5167).

The following significant (semver-major) changes have been made since the previous Node v5.0.0 release. * Buffer * New Buffer constructors have been added [#4682](#4682) * Previously deprecated Buffer APIs are removed [#5048](#5048), [#4594](#4594) * Improved error handling [#4514](#4514) * Cluster * Worker emitted as first argument in 'message' event [#5361](#5361). * Crypto * Improved error handling [#3100](#3100), [#5611](#5611) * Simplified Certificate class bindings [#5382](#5382) * Improved control over FIPS mode [#5181](#5181) * pbkdf2 digest overloading is deprecated [#4047](#4047) * Dependencies * Reintroduce shared c-ares build support [#5775](#5775). * V8 updated to 5.0.71.31 [#6111](#6111). * DNS * Add resolvePtr API to query plain DNS PTR records [#4921](#4921). * Domains * Clear stack when no error handler [#4659](#4659). * File System * The `fs.realpath()` and `fs.realpathSync()` methods have been updated to use a more efficient libuv implementation. This change includes the removal of the `cache` argument and the method can throw new errors [#3594](#3594) * FS apis can now accept and return paths as Buffers [#5616](#5616). * Error handling and type checking improvements [#5616](#5616), [#5590](#5590), [#4518](#4518), [#3917](#3917). * fs.read's string interface is deprecated [#4525](#4525) * HTTP * 'clientError' can now be used to return custom errors from an HTTP server [#4557](#4557). * Modules * Current directory is now prioritized for local lookups [#5689](#5689) * Symbolic links are preserved when requiring modules [#5950](#5950) * Net * DNS hints no longer implicitly set [#6021](#6021). * Improved error handling and type checking [#5981](#5981), [#5733](#5733), [#2904](#2904) * OS X * MACOSX_DEPLOYMENT_TARGET has been bumped up to 10.7 [#6402](#6402). * Path * Improved type checking [#5348](#5348). * Process * Introduce process warnings API [#4782](#4782). * Throw exception when non-function passed to nextTick [#3860](#3860). * Readline * Emit key info unconditionally [#6024](#6024) * REPL * Assignment to `_` will emit a warning. [#5535](#5535) * Timers * Fail early when callback is not a function [#4362](#4362) * TLS * Rename 'clientError' to 'tlsClientError' [#4557](#4557) * SHA1 used for sessionIdContext [#3866](#3866) * TTY * Previously deprecated setRawMode wrapper is removed [#2528](#2528). * Util * Changes to Error object formatting [#4582](#4582). * Windows * Windows XP and Vista are no longer supported [#5167](#5167), [#5167](#5167).

feross reviewed Jan 3, 2016
View changes

feross mentioned this pull request Jan 3, 2016

fs.read into zero buffer should not throw exception #4517

Closed

thefourtheye reviewed Jan 3, 2016
View changes

thefourtheye added the fs label Jan 3, 2016

thefourtheye reviewed Jan 4, 2016
View changes

feross mentioned this pull request Jan 4, 2016

Deprecate legacy fs.read and fs.readSync arguments #4530

Closed

thefourtheye reviewed Jan 4, 2016
View changes

jasnell added the semver-major label Jan 5, 2016

jasnell reviewed Jan 6, 2016
View changes

fs.read into zero buffer should not throw exception

8c0ad82

Fixes #4517.

jasnell added a commit that referenced this pull request Jan 6, 2016

fs: fs.read into zero buffer should not throw exception

2b15e68

Fixes: #4517. PR-URL: #4518 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com>

jasnell closed this Jan 6, 2016

feross deleted the feross:fs-read-fix branch Jan 6, 2016

jasnell mentioned this pull request Mar 17, 2016

Planning for v6 #5766

Closed

scovetta pushed a commit to scovetta/node that referenced this pull request Apr 2, 2016

fs: fs.read into zero buffer should not throw exception

b500706

Fixes: nodejs#4517. PR-URL: nodejs#4518 Reviewed-By: Evan Lucas <evanlucas@me.com> Reviewed-By: James M Snell <jasnell@gmail.com>

jasnell mentioned this pull request Apr 19, 2016

What is new in v6? #6264

Closed

jasnell mentioned this pull request Apr 26, 2016

2016-04-26, Version 6.0.0 (Current) Release #6383

Closed

nodejs / node

fs.read into zero buffer should not throw exception #4518

fs.read into zero buffer should not throw exception #4518

Conversation

feross commented Jan 3, 2016

This comment has been minimized.

feross Jan 3, 2016 Author Contributor

This comment has been minimized.

bnoordhuis Jan 3, 2016 Member

This comment has been minimized.

feross Jan 3, 2016 Author Contributor

Trott commented Jan 3, 2016

This comment has been minimized.

thefourtheye Jan 3, 2016 Contributor

This comment has been minimized.

feross Jan 3, 2016 Author Contributor

thefourtheye commented Jan 3, 2016

Trott commented Jan 3, 2016

feross commented Jan 3, 2016

feross commented Jan 3, 2016

tejasmanohar commented Jan 3, 2016

Trott commented Jan 3, 2016

feross commented Jan 3, 2016

Trott commented Jan 3, 2016

evanlucas commented Jan 4, 2016

This comment has been minimized.

thefourtheye Jan 4, 2016 Contributor

This comment has been minimized.

feross Jan 4, 2016 Author Contributor

This comment has been minimized.

feross Jan 4, 2016 Author Contributor

feross commented Jan 4, 2016

This comment has been minimized.

thefourtheye Jan 4, 2016 Contributor

This comment has been minimized.

feross Jan 6, 2016 Author Contributor

This comment has been minimized.

thefourtheye Jan 4, 2016 Contributor

This comment has been minimized.

feross Jan 5, 2016 Author Contributor

This comment has been minimized.

trevnorris Jan 6, 2016 Contributor

This comment has been minimized.

thefourtheye Jan 6, 2016 Contributor

jasnell commented Jan 5, 2016

feross commented Jan 6, 2016

thefourtheye commented Jan 6, 2016

This comment has been minimized.

jasnell Jan 6, 2016 Member

This comment has been minimized.

feross Jan 6, 2016 Author Contributor

This comment has been minimized.

Fishrock123 Jan 6, 2016 Member

This comment has been minimized.

jasnell Jan 6, 2016 Member

This comment has been minimized.

feross Jan 6, 2016 Author Contributor

This comment has been minimized.

jasnell Jan 6, 2016 Member

This comment has been minimized.

feross Jan 6, 2016 Author Contributor

feross commented Jan 6, 2016

jasnell commented Jan 6, 2016

feross commented Jan 6, 2016

jasnell commented Jan 6, 2016

jasnell commented Jan 6, 2016

feross commented Jan 6, 2016

feross Jan 3, 2016
Author Contributor

bnoordhuis Jan 3, 2016
Member

feross Jan 3, 2016
Author Contributor

thefourtheye Jan 3, 2016
Contributor

feross Jan 3, 2016
Author Contributor

thefourtheye Jan 4, 2016
Contributor

feross Jan 4, 2016
Author Contributor

feross Jan 4, 2016
Author Contributor

thefourtheye Jan 4, 2016
Contributor

feross Jan 6, 2016
Author Contributor

thefourtheye Jan 4, 2016
Contributor

feross Jan 5, 2016
Author Contributor

trevnorris Jan 6, 2016
Contributor

thefourtheye Jan 6, 2016
Contributor

jasnell Jan 6, 2016
Member

feross Jan 6, 2016
Author Contributor

Fishrock123 Jan 6, 2016
Member

jasnell Jan 6, 2016
Member

feross Jan 6, 2016
Author Contributor

jasnell Jan 6, 2016
Member

feross Jan 6, 2016
Author Contributor