-
Notifications
You must be signed in to change notification settings - Fork 11.3k
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
[5.8] Strip slashes for unique rule helper #27940
Conversation
@@ -50,8 +50,8 @@ public function ignore($id, $idColumn = null) | |||
*/ | |||
public function ignoreModel($model, $idColumn = null) | |||
{ | |||
$this->idColumn = str_replace(',', '', $idColumn ?? $model->getKeyName()); | |||
$this->ignore = str_replace(',', '', $model->{$this->idColumn}); | |||
$this->idColumn = addslashes($idColumn ?? $model->getKeyName()); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why would addslashes
be relevant for the column, which I would expect is never a user provided input?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm I wonder why a “user” would request anything with any special characters.... maybe because they are maliciously trying things to a site?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think mfn's point is the same as the one Taylor made in the blog post: the developer using Laravel should never expose the ability for their user to define the database column through an input.
Worst case scenario the user input should be decoupled from the database, which would never be vulnerable.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this could have simple avoided by type-hinting the method
@@ -35,8 +35,8 @@ public function ignore($id, $idColumn = null) | |||
return $this->ignoreModel($id, $idColumn); | |||
} | |||
|
|||
$this->ignore = str_replace(',', '', $id); | |||
$this->idColumn = str_replace(',', '', $idColumn ?? 'id'); | |||
$this->ignore = addslashes($id); |
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.
Stupid q: is this escaping so the value can be literally used in a DB query, where no further escaping happens?
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.
Using addslashes()
to escape a value for a DB query is quite dangerous.
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.
My thoughts, thus I'm asking. Not very familiar with the exact code, though.
But I know at least for Postgres it doesn't even work correctly.
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.
We aren't adding slashes for the DB query. We are adding slashes for the str_getcsv
call in the validator. This has nothing to do with trying to escape data for a DB query.
To clarify, the vulnerability is that a user can construct a string that is sent through the str_getcsv call by the validator and uses a column of their choosing in the DB query. The value is bound as a PDO parameter. However, columns are not. Therefore, we need to ensure that the "id" portion of the unique validation rule is a single string and can't be manipulated with commas / quotes to create multiple values when parsed by str_getcsv.
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.
👍 thanks for clarification.
I think however that the code could use some comments as to why all these special handling.
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.
Because a vulnerability was reported. It was possible to perform a XSS using the ID
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.
It's not XSS.
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 know this is old article but i guess it tells you why addslashes is not perfect solution here - http://shiflett.org/blog/2006/addslashes-versus-mysql-real-escape-string Not sure though how it's realted to what @taylorotwell wrote, i guess someone needs to check it out.
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.
Keep in mind that Taylor already clarified, this is not for escaping for the database but for another case.
Also, the code already changed in the meantime and the addslashes
was moved to the __toString
and only for the ignore
and not the column.
A regression test would be cool. 😉 |
Good job guys! You saved probably a trillion dollars by acting so fast on this to many organizations who wouldn't even had known! If i was filthy rich I'd buy each of you |
This removes the need to introduce a breaking change while prevents manipulating the generated rule string.