332472 - new RegExp() ignores string boundaries when throwing exceptions

Status: VERIFIED FIXED

(Core :: JavaScript Engine, defect, P1)

Description

[Wladimir Palant]

The issue is - the exception thrown by new RegExp() shows more text than it should. JavaScript code to test this:

var str1 = "?asdf\nAnd you really shouldn't see this!";
var str2 = str1.substr(0, 5);
try {
  new RegExp(str2);
}
catch(e) {
  alert(e);
}

This should only display: "invalid qualifier ?asdf". Instead it displays "And you really shouldn't see this" as well - a string that isn't even in str2! So obviously the string length is ignored when the exception is created, instead everything until the next \0 is used.

I tested this in rv:1.8.0.1 Gecko/20060111 Firefox/1.5.0.1 and in rv:1.9a1 Gecko/20060309 Firefox/1.6a1, results are the same in both.

I am marking this bug security relevant until somebody can confirm that you can't get access to random memory blocks that way.

Comment 1

[Wladimir Palant]

Attachment: Testcase

Comment 2

[Brendan Eich [:brendan]]

Attachment: proposed fix

Seems best to use an independent string for compiling, and retain it as the value of the source property.

/be

Comment 3

[Brendan Eich [:brendan]]

This is not security sensitive.  JS implements dependent strings, so e.g., "hello, world".substr(0, 4) is a string header that points into the "hello, world" string's buffer.  The dependent string optimization is thread- and memory-safe.  It's just hazardous when pointers to char substrings flow into bad old code that expects to find a NUL terminator.  The independent string ("hello, world" in the example) that dependent strings share *always* ends with a backstop NUL at index equal to length.

/be

Comment 4

[Jesse Ruderman]

See also bug 330796, "Invalid quantifier" error message shows wrong part of regexp.

Comment 5

[Brendan Eich [:brendan]]

(In reply to comment #4)
> See also bug 330796, "Invalid quantifier" error message shows wrong part of
> regexp.

Unrelated.  See bug 330796 comment 3.

/be

Comment 6

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 216950 [details] [diff] [review]
proposed fix

r=mrbkap

Comment 7

[Brendan Eich [:brendan]]

Fixed on trunk and 1.8 branch.

/be

Comment 8

[Bob Clary [:bc] (inactive)]

Checking in regress-332472.js;
/cvsroot/mozilla/js/tests/js1_5/Exceptions/regress-332472.js,v  <--  regress-332472.js
initial revision: 1.1

Comment 9

[Bob Clary [:bc] (inactive)]

verified fixed 1.8.1/1.9a1 win/mac/linux 20060415

Comment 10

[Bob Clary [:bc] (inactive)]

not on 1.8.0, not making js16