-
Notifications
You must be signed in to change notification settings - Fork 30k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
buffer: throw on failed fill attempts #17427
Conversation
src/node_buffer.cc
Outdated
if (str_length == 0) | ||
return; | ||
return env->ThrowError("Buffer could not be filled"); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not super familiar with this code but could this return a non-undefined value (that is, not from this return
statement but the usual args.GetReturnValue().Set()
) and then throw in JS? Then we could have an error code for it and document it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Throwing from JavaScript would be ideal
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like nodejs#17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. Refs: nodejs#17427 Refs: nodejs#17423
I'm good with this but the new error needs to have an internal/errors error code assigned. |
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like #17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: #17428 Refs: #17427 Refs: #17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like nodejs#17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: nodejs#17428 Refs: nodejs#17427 Refs: nodejs#17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
7209946
to
8b7ad01
Compare
Updated to throw in JavaScript. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The documentation in buffer.md
has to be updated as well and it would be good to switch to a TypeError.
Besides that this is the right approach out of my perspective and LGTM and we should do this, even if it is a BC.
lib/buffer.js
Outdated
|
||
return bindingFill(buf, val, start, end, encoding); | ||
if (fillLength === -1) | ||
throw new errors.Error('ERR_INVALID_ARG_VALUE', 'value', val); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This should be a TypeError
as far as I see it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
+1... TypeError
would be better here.
Oh, it seems the alternative already landed. I personally think this is actually the better approach but as that was already decided, I am closing this. |
@BridgeAR I think this was still open because this might be the direction for 10.x while the other PR takes care of all other active release lines. I think it should be re-opened potentially? |
|
Thanks, seems like I did not look closely enough in the other one. The documentation needs another update though. There is a example that would throw now in |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM if changed to use TypeError
Updated to a |
If fill() attempts to write a string to a buffer, but fails silently, then uninitialized memory could be leaked. This commit causes fill() to throw if the string write operation fails. Refs: nodejs#17423 PR-URL: nodejs#17427 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Ruben Bridgewater <[email protected]>
@cjihrig even though I guess this is a uncontroversial commit, it did not get two TSC approvals yet. |
|
||
```js | ||
const buf = Buffer.allocUnsafe(5); | ||
// Prints: <Buffer 61 61 61 61 61> | ||
console.log(buf.fill('a')); | ||
// Prints: <Buffer aa aa aa aa aa> | ||
console.log(buf.fill('aazz', 'hex')); | ||
// Prints: <Buffer aa aa aa aa aa> | ||
// Throws a exception. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Micro-nit: a
-> an
Heh, we were just talking about stuff like this (sort of) in Build WG yesterday. One more data point/argument for moving to more automation. @maclover7 |
I'd like to propose dropping that rule. Or at least provide some type of time boxing where it no longer applies if there are no objections. |
Sounds like re-opening nodejs/TSC#378. I strongly opposed such a move, although I was in the minority. I still oppose it although I am probably still in the minority. If two TSC members (out of 20!!!) can't be moved to review a change in a timely fashion, fix the TSC. Our first impulse is always to accommodate disengaged folks. We need to stop that. In fairness to TSC folks, there wasn't any ping on this one so it may have slipped under the radar for lots of folks. If we want to change the rules for semver-major changes, let's make that change: Require @-mentioning @nodejs/tsc. |
Requiring an @ mention sounds fair enough. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
After-the-fact LGTM.
Refs: nodejs#17427 PR-URL: nodejs#17501 Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Rich Trott <[email protected]>
rather than @ mentioning I find adding @nodejs/tsc as a reviewer is a much better approach, that way it ends up in peoples github list of issues to review (as well as pinging them) |
That could be both a requirement. |
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like #17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: #17428 Backport-PR-URL: #17467 Refs: #17427 Refs: #17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like #17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: #17428 Backport-PR-URL: #17467 Refs: #17427 Refs: #17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like #17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: #17428 Refs: #17427 Refs: #17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
Zero-fill when `Buffer.alloc()` receives invalid fill data. A solution like #17427 which switches to throwing makes sense, but is likely a breaking change. This suggestion leaves the behaviour of `buffer.fill()` untouched, since any change to it would be a breaking change, and lets `Buffer.alloc()` check whether any filling took place or not. PR-URL: #17428 Refs: #17427 Refs: #17423 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Colin Ihrig <[email protected]> Reviewed-By: Anatoli Papirovski <[email protected]> Reviewed-By: Ali Ijaz Sheikh <[email protected]> Reviewed-By: Tobias Nießen <[email protected]> Reviewed-By: Luigi Pinca <[email protected]> Reviewed-By: Michaël Zasso <[email protected]>
If
fill()
attempts to write a string to a buffer, but fails silently, then uninitialized memory could be leaked. This commit causesfill()
to throw if the string write operation fails.This also addresses an existing TODO comment.
Fixes: #17423
Checklist
make -j4 test
(UNIX), orvcbuild test
(Windows) passesAffected core subsystem(s)
buffer