Is the following javascript safe from arbitrary code execution? - javascript

I'm contributing to a javascript framework which has the equivalent of the following code:
eval("'" + user_input.replace(/'/g, "'") + "'");
I know this is terrible -- no need to persuade me. What I want to know is, can I inject arbitrary code here?
At first glance the user_input.replace("'", "'") would seem to prevent me from breaking out of the string. However I can pass in newlines e.g. \nalert(123)\n, but then the result is always a syntax error, e.g.
'
alert(123)
'
Is there actually a vector for code injection here, other than just causing a syntax error?

While this is undoubtedly a worrisome pattern, it's safe if used exactly in the way described. The only character that can terminate a single-quoted string in Javascript is the single quote character. So long as that character does not appear in the string interpolated into the single quotes, it cannot possibly be interpreted as anything other than a string.
About the worst thing I can think of that you could do is end a string with a backslash, which would result in an unterminated string, e.g. if user_input were:
example\
then the evaluated code would be
'example\'
which would result in a syntax error, because the string contained in the eval is never terminated. However, if the real eval is actually more complex, this is exploitable. For example, if the code were:
var escaped_input = user_input.replace(/'/g, "&39;");
eval("'" + escaped_input + "' some more stuff '" + escaped_input + "'");
then it could be exploited with an input like:
; alert(1); // \
which would result in:
'; alert(1); // \' some more stuff '; alert(1); // \'
^^^^^^^^^
in which the underlined content would be evaluated, because the quote that was supposed to exit the string was escaped, turning the next single quote into a closing quote! To be safe, I'd recommend escaping or replacing backslashes if possible (unless you're explicitly trying to use eval() to deal with them, in which case you might just catch the exception).

Related

A way to automatically enclosure quotes inside string containing code (RegExp maybe?)

Is there a way to automatically enclosure all quotes inside the string of js-code which contains other strings by itself? I just failed to make a RegExp valuable to identify all the quotes properly. Because in some cases (blocks of code generated by script) there strings of code like:
$(".class1.class2:contains('text')").param(s1+s2+"-"+s3+"_"+i)...
var r=new RegExp("blabla[qwer]\\'\\w+\\d\\'");
//etc.
More complex actually and less readable =)
But anyways is there a way to get rid of constant http/ajax/json requests and make code to work without injection but via eval? It's not intended to be used in outer net so it's safe, dont tell me anything 'bout eval =) The main problem: how to make enclosurement of quotes inside the string...
Edit:
Sorry i just don't know how to say it in English, so by enclosurement i mean insertion of a specific symbol before char which has dual meaning to clarify it's definition for compiler or machine-interpreter.
Example1: ('first_part_of_string enclosured_quote=\' second_part_of_string')
Example2: `<- these symbols don't get interpreted as code-markers because of enclosurement ->`
Sure, here it is with single quotes:
'
$(".class1.class2:contains(\'text\')").param(s1+s2+"-"+s3+"_"+i)...
var r=new RegExp("blabla[qwer]\\\\\'\\\w+\\\d\\\\\'");
'
here it is with double quotes:
"
$(\".class1.class2:contains('text')\").param(s1+s2+\"-\"+s3+\"_\"+i)...
var r=new RegExp(\"blabla[qwer]\\\\'\\\\w+\\\\d\\\\'\");
"
You can use the Mega-String tool to do this automatically
I used the single / double quote option from the Other category..

What is the difference of operators " and ' in javascript [duplicate]

Consider the following two alternatives:
console.log("double");
console.log('single');
The former uses double quotes around the string, whereas the latter uses single quotes around the string.
I see more and more JavaScript libraries out there using single quotes when handling strings.
Are these two usages interchangeable? If not, is there an advantage in using one over the other?
The most likely reason for use of single vs. double in different libraries is programmer preference and/or API consistency. Other than being consistent, use whichever best suits the string.
Using the other type of quote as a literal:
alert('Say "Hello"');
alert("Say 'Hello'");
This can get complicated:
alert("It's \"game\" time.");
alert('It\'s "game" time.');
Another option, new in ECMAScript 6, is template literals which use the backtick character:
alert(`Use "double" and 'single' quotes in the same string`);
alert(`Escape the \` back-tick character and the \${ dollar-brace sequence in a string`);
Template literals offer a clean syntax for: variable interpolation, multi-line strings, and more.
Note that JSON is formally specified to use double quotes, which may be worth considering depending on system requirements.
If you're dealing with JSON, it should be noted that strictly speaking, JSON strings must be double quoted. Sure, many libraries support single quotes as well, but I had great problems in one of my projects before realizing that single quoting a string is in fact not according to JSON standards.
There is no one better solution; however, I would like to argue that double quotes may be more desirable at times:
Newcomers will already be familiar with double quotes from their language. In English, we must use double quotes " to identify a passage of quoted text. If we were to use a single quote ', the reader may misinterpret it as a contraction. The other meaning of a passage of text surrounded by the ' indicates the 'colloquial' meaning. It makes sense to stay consistent with pre-existing languages, and this may likely ease the learning and interpretation of code.
Double quotes eliminate the need to escape apostrophes (as in contractions). Consider the string: "I'm going to the mall", vs. the otherwise escaped version: 'I\'m going to the mall'.
Double quotes mean a string in many other languages. When you learn a new language like Java or C, double quotes are always used. In Ruby, PHP and Perl, single-quoted strings imply no backslash escapes while double quotes support them.
JSON notation is written with double quotes.
Nonetheless, as others have stated, it is most important to remain consistent.
Section 7.8.4 of the specification describes literal string notation. The only difference is that DoubleStringCharacter is "SourceCharacter but not double-quote" and SingleStringCharacter is "SourceCharacter but not single-quote". So the only difference can be demonstrated thusly:
'A string that\'s single quoted'
"A string that's double quoted"
So it depends on how much quote escaping you want to do. Obviously the same applies to double quotes in double quoted strings.
I'd like to say the difference is purely stylistic, but I'm really having my doubts. Consider the following example:
/*
Add trim() functionality to JavaScript...
1. By extending the String prototype
2. By creating a 'stand-alone' function
This is just to demonstrate results are the same in both cases.
*/
// Extend the String prototype with a trim() method
String.prototype.trim = function() {
return this.replace(/^\s+|\s+$/g, '');
};
// 'Stand-alone' trim() function
function trim(str) {
return str.replace(/^\s+|\s+$/g, '');
};
document.writeln(String.prototype.trim);
document.writeln(trim);
In Safari, Chrome, Opera, and Internet Explorer (tested in Internet Explorer 7 and Internet Explorer 8), this will return the following:
function () {
return this.replace(/^\s+|\s+$/g, '');
}
function trim(str) {
return str.replace(/^\s+|\s+$/g, '');
}
However, Firefox will yield a slightly different result:
function () {
return this.replace(/^\s+|\s+$/g, "");
}
function trim(str) {
return str.replace(/^\s+|\s+$/g, "");
}
The single quotes have been replaced by double quotes. (Also note how the indenting space was replaced by four spaces.) This gives the impression that at least one browser parses JavaScript internally as if everything was written using double quotes. One might think, it takes Firefox less time to parse JavaScript if everything is already written according to this 'standard'.
Which, by the way, makes me a very sad panda, since I think single quotes look much nicer in code. Plus, in other programming languages, they're usually faster to use than double quotes, so it would only make sense if the same applied to JavaScript.
Conclusion: I think we need to do more research on this.
This might explain Peter-Paul Koch's test results from back in 2003.
It seems that single quotes are sometimes faster in Explorer Windows (roughly 1/3 of my tests did show a faster response time), but if Mozilla shows a difference at all, it handles double quotes slightly faster. I found no difference at all in Opera.
2014: Modern versions of Firefox/Spidermonkey don’t do this anymore.
If you're doing inline JavaScript (arguably a "bad" thing, but avoiding that discussion) single quotes are your only option for string literals, I believe.
E.g., this works fine:
<a onclick="alert('hi');">hi</a>
But you can't wrap the "hi" in double quotes, via any escaping method I'm aware of. Even " which would have been my best guess (since you're escaping quotes in an attribute value of HTML) doesn't work for me in Firefox. " won't work either because at this point you're escaping for HTML, not JavaScript.
So, if the name of the game is consistency, and you're going to do some inline JavaScript in parts of your application, I think single quotes are the winner. Someone please correct me if I'm wrong though.
Technically there's no difference. It's only matter of style and convention.
Douglas Crockford1 recommends using double quotes.
I personally follow that.
Strictly speaking, there is no difference in meaning; so the choice comes down to convenience.
Here are several factors that could influence your choice:
House style: Some groups of developers already use one convention or the other.
Client-side requirements: Will you be using quotes within the strings? (See Ady's answer.)
Server-side language: VB.NET people might choose to use single quotes for JavaScript so that the scripts can be built server-side (VB.NET uses double-quotes for strings, so the JavaScript strings are easy to distinguished if they use single quotes).
Library code: If you're using a library that uses a particular style, you might consider using the same style yourself.
Personal preference: You might think one or other style looks better.
Just keep consistency in what you use. But don't let down your comfort level.
"This is my string."; // :-|
"I'm invincible."; // Comfortable :)
'You can\'t beat me.'; // Uncomfortable :(
'Oh! Yes. I can "beat" you.'; // Comfortable :)
"Do you really think, you can \"beat\" me?"; // Uncomfortable :(
"You're my guest. I can \"beat\" you."; // Sometimes, you've to :P
'You\'re my guest too. I can "beat" you too.'; // Sometimes, you've to :P
ECMAScript 6 update
Using template literal syntax.
`Be "my" guest. You're in complete freedom.`; // Most comfort :D
I hope I am not adding something obvious, but I have been struggling with Django, Ajax, and JSON on this.
Assuming that in your HTML code you do use double quotes, as normally should be, I highly suggest to use single quotes for the rest in JavaScript.
So I agree with ady, but with some care.
My bottom line is:
In JavaScript it probably doesn't matter, but as soon as you embed it inside HTML or the like you start to get troubles. You should know what is actually escaping, reading, passing your string.
My simple case was:
tbox.innerHTML = tbox.innerHTML + '<div class="thisbox_des" style="width:210px;" onmouseout="clear()"><a href="/this/thislist/'
+ myThis[i].pk +'"><img src="/site_media/'
+ myThis[i].fields.thumbnail +'" height="80" width="80" style="float:left;" onmouseover="showThis('
+ myThis[i].fields.left +','
+ myThis[i].fields.right +',\''
+ myThis[i].fields.title +'\')"></a><p style="float:left;width:130px;height:80px;"><b>'
+ myThis[i].fields.title +'</b> '
+ myThis[i].fields.description +'</p></div>'
You can spot the ' in the third field of showThis.
The double quote didn't work!
It is clear why, but it is also clear why we should stick to single quotes... I guess...
This case is a very simple HTML embedding, and the error was generated by a simple copy/paste from a 'double quoted' JavaScript code.
So to answer the question:
Try to use single quotes while within HTML. It might save a couple of debug issues...
It's mostly a matter of style and preference. There are some rather interesting and useful technical explorations in the other answers, so perhaps the only thing I might add is to offer a little worldly advice.
If you're coding in a company or team, then it's probably a good idea to follow the "house style".
If you're alone hacking a few side projects, then look at a few prominent leaders in the community. For example, let's say you getting into Node.js. Take a look at core modules, for example, Underscore.js or express and see what convention they use, and consider following that.
If both conventions are equally used, then defer to your personal preference.
If you don't have any personal preference, then flip a coin.
If you don't have a coin, then beer is on me ;)
I am not sure if this is relevant in today's world, but double quotes used to be used for content that needed to have control characters processed and single quotes for strings that didn't.
The compiler will run string manipulation on a double quoted string while leaving a single quoted string literally untouched. This used to lead to 'good' developers choosing to use single quotes for strings that didn't contain control characters like \n or \0 (not processed within single quotes) and double quotes when they needed the string parsed (at a slight cost in CPU cycles for processing the string).
If you are using JSHint, it will raise an error if you use a double quoted string.
I used it through the Yeoman scafflholding of AngularJS, but maybe there is somehow a manner to configure this.
By the way, when you handle HTML into JavaScript, it's easier to use single quote:
var foo = '<div class="cool-stuff">Cool content</div>';
And at least JSON is using double quotes to represent strings.
There isn't any trivial way to answer to your question.
There isn't any difference between single and double quotes in JavaScript.
The specification is important:
Maybe there are performance differences, but they are absolutely minimum and can change any day according to browsers' implementation. Further discussion is futile unless your JavaScript application is hundreds of thousands lines long.
It's like a benchmark if
a=b;
is faster than
a = b;
(extra spaces)
today, in a particular browser and platform, etc.
Examining the pros and cons
In favor of single quotes
Less visual clutter.
Generating HTML: HTML attributes are usually delimited by double quotes.
elem.innerHTML = 'Hello';
However, single quotes are just as legal in HTML.
elem.innerHTML = "<a href='" + url + "'>Hello</a>";
Furthermore, inline HTML is normally an anti-pattern. Prefer templates.
Generating JSON: Only double quotes are allowed in JSON.
myJson = '{ "hello world": true }';
Again, you shouldn’t have to construct JSON this way. JSON.stringify() is often enough. If not, use templates.
In favor of double quotes
Doubles are easier to spot if you don't have color coding. Like in a console log or some kind of view-source setup.
Similarity to other languages: In shell programming (Bash etc.), single-quoted string literals exist, but escapes are not interpreted inside them. C and Java use double quotes for strings and single quotes for characters.
If you want code to be valid JSON, you need to use double quotes.
In favor of both
There is no difference between the two in JavaScript. Therefore, you can use whatever is convenient at the moment. For example, the following string literals all produce the same string:
"He said: \"Let's go!\""
'He said: "Let\'s go!"'
"He said: \"Let\'s go!\""
'He said: \"Let\'s go!\"'
Single quotes for internal strings and double for external. That allows you to distinguish internal constants from strings that are to be displayed to the user (or written to disk etc.). Obviously, you should avoid putting the latter in your code, but that can’t always be done.
Talking about performance, quotes will never be your bottleneck. However, the performance is the same in both cases.
Talking about coding speed, if you use ' for delimiting a string, you will need to escape " quotes. You are more likely to need to use " inside the string. Example:
// JSON Objects:
var jsonObject = '{"foo":"bar"}';
// HTML attributes:
document.getElementById("foobar").innerHTML = '<input type="text">';
Then, I prefer to use ' for delimiting the string, so I have to escape fewer characters.
There are people that claim to see performance differences: old mailing list thread. But I couldn't find any of them to be confirmed.
The main thing is to look at what kind of quotes (double or single) you are using inside your string. It helps to keep the number of escapes low. For instance, when you are working with HTML content inside your strings, it is easier to use single quotes so that you don't have to escape all double quotes around the attributes.
When using CoffeeScript I use double quotes. I agree that you should pick either one and stick to it. CoffeeScript gives you interpolation when using the double quotes.
"This is my #{name}"
ECMAScript 6 is using back ticks (`) for template strings. Which probably has a good reason, but when coding, it can be cumbersome to change the string literals character from quotes or double quotes to backticks in order to get the interpolation feature. CoffeeScript might not be perfect, but using the same string literals character everywhere (double quotes) and always be able to interpolate is a nice feature.
`This is my ${name}`
I would use double quotes when single quotes cannot be used and vice versa:
"'" + singleQuotedValue + "'"
'"' + doubleQuotedValue + '"'
Instead of:
'\'' + singleQuotedValue + '\''
"\"" + doubleQuotedValue + "\""
There is strictly no difference, so it is mostly a matter of taste and of what is in the string (or if the JavaScript code itself is in a string), to keep number of escapes low.
The speed difference legend might come from PHP world, where the two quotes have different behavior.
The difference is purely stylistic. I used to be a double-quote Nazi. Now I use single quotes in nearly all cases. There's no practical difference beyond how your editor highlights the syntax.
You can use single quotes or double quotes.
This enables you for example to easily nest JavaScript content inside HTML attributes, without the need to escape the quotes.
The same is when you create JavaScript with PHP.
The general idea is: if it is possible use such quotes that you won't need to escape.
Less escaping = better code.
In addition, it seems the specification (currently mentioned at MDN) doesn't state any difference between single and double quotes except closing and some unescaped few characters. However, template literal (` - the backtick character) assumes additional parsing/processing.
A string literal is 0 or more Unicode code points enclosed in single or double quotes. Unicode code points may also be represented by an escape sequence. All code points may appear literally in a string literal except for the closing quote code points, U+005C (REVERSE SOLIDUS), U+000D (CARRIAGE RETURN), and U+000A (LINE FEED). Any code points may appear in the form of an escape sequence. String literals evaluate to ECMAScript String values...
Source: https://tc39.es/ecma262/#sec-literals-string-literals

Error due to single quote ' in javascript function call

I am passing value dynamically to javascript function.
I am retrieving data from database and filling to javascript function, it does not have a static binding.
share_it(data_from_mysql_database);
like
share_it('value from mysql database');
Some times value contain a single quote (').
like:
share_it(' Essentially you'll have to have a good academic history ');
So function call gives error that:
Uncaught SyntaxError: Unexpected identifier
You can use the \ character to escape such characters:
share_it(' Essentially you\'ll have to have a good past academic ');
Or, you can switch to using double quotes if you know you will need to embed a single quote character:
share_it(" Essentially you'll have to have a good past academic ");
You can freely switch between double " and single ' quotes where you need the other in a literal string:
share_it(" Essentially you'll have to have a good past academic ");
Only in cases where you need both, you need to escape the repeating character:
share_it(" Essentially you'll have to have a good \"past\" academic ");
You can also replace the ' in the string with &#39.
You ought to be converting special chars on the upstream rather than the downstream. Converting it on the upstream saves time later when an inexperienced developer does not care to escape the data on the downstream when sent to the client. Since you have not properly converted the data on the upstream, you have no choice. You should escape it.
share_it(escape(data_from_mysql_database));
Example"
> escape("You're awesome");
'You%27re%20awesome'
>

What does ' ', and " ", and no quotes mean in Javascript?

I realized I've been switching between them with no understanding as to why, and am finding it hard to search for.
' ' and " " are the same thing; they are used to define string literals.
Things without quotes can be an identifier, keyword, non-string literal, property name or a number (may have missed one).
Examples:
"hello world" literal (string)
'hello world' literal (string) with same contents
document identifier (object)
{ a: 1 } property name
if keyword (start conditional statement)
3.4 literal (number)
/abc/ literal (regex object)
String literals that are enclosed in single quotes don't need escaped double quotes and visa versa, e.g.:
'click me' HTML containing double quotes
"It's going to rain" String containing single quote
' ' and " " used to quote string literal and represents string(s) whereas literal without quote are variables (name of variable, constant) know as identifier, example
variable = 'Hello'; (Here `variable` is identifier and 'Hello' is string literal)
var = "Ho There"
You might question, what is the difference between ' (single quote) and " (Double quote)
Difference is ,strings within " if have special character then they need to escape. Example:
Variable = "hi " there"; ---> here you need to escape the " inside string like
Variable = "hi \" there";
But if using, ' then no need of escaping (unless there is a extra ' in string). You can hve like
var = 'Hello " World"';
" and ' are interchangeable (but need to be used together).
myObject["property"] and myObject.property are also interchangeable. $var foo = "property"; myObject[foo] as well (per comment below).
A quick jsfiddle around and both single and double quotes escape control codes etc.
In latter days I have had errors from HTML where double quotes have not been used, and if you look at the spec for JSON you'll note it is a double quote that is asked for when quoting string literals. So it is double quotes that is the convention I think, for historical reasons.
However! In these days of writing server side JS I must admit I tend to be pulled back to my C roots and double quote where I want escaped chars and single quote strings that are effectively literal chars and never likely to contain escaped chars (even though this is essentially non-productive behaviour). Besides which most of my JS is coffeescript nowadays anyway, nobody ever wrote javascript for elegance, CS is a different kettle of fish though.

Double-Escaped Unicode Javascript Issue

I am having a problem displaying a Javascript string with embedded Unicode character escape sequences (\uXXXX) where the initial "\" character is itself escaped as "\"
What do I need to do to transform the string so that it properly evaluates the escape sequences and produces output with the correct Unicode character?
For example, I am dealing with input such as:
"this is a \u201ctest\u201d";
attempting to decode the "\" using a regex expression, e.g.:
var out = text.replace('/\/g','\');
results in the output text:
"this is a \u201ctest\u201d";
that is, the Unicode escape sequences are displayed as actual escape sequences, not the double quote characters I would like.
As it turns out, it's unescape() we want, but with '%uXXXX' rather than '\uXXXX':
unescape(yourteststringhere.replace(/\/g,'%'))
This is a terrible solution, but you can do this:
var x = "this is a \u201ctest\u201d".replace(/\/g,'\\')
// x is now "this is a \u201ctest\u201d"
eval('x = "' + x + '"')
// x is now "this is a “test”"
It's terrible because:
eval can be dangerous, if you don't know what's in the string
the string quoting in the eval statement will break if you have actual quotation marks in your string
Are you sure '\' is the only character that might get HTML-escaped? Are you sure '\uXXXX' is the only kind of string escape in use?
If not, you'll need a general-purpose HTML-character/entity-reference-decoder and JS-string-literal-decoder. Unfortunately JavaScript has no built-in methods for this and it's quite tedious to do manually with a load of regexps.
It is possible to take advantage of the browser's HTML-decoder by assigning the string to an element's innerHTML property, and then ask JavaScript to decode the string as above:
var el= document.createElement('div');
el.innerHTML= s;
return eval('"'+el.firstChild.data+'"');
However this is an incredibly ugly hack and a security hole if the string comes from a source that isn't 100% trusted.
Where are the strings coming from? It would be nicer if possible to deal with the problem at the server end where you may have more powerful text handling features available. And if you could fix whatever it is that is unnecessarily HTML-escaping your backslashes you could find the problem fixes itself.
I'm not sure if this is it, but the answer might have something to do with eval(), if you can trust your input.
I was thinking along the same lines, but using eval() in everyway I could imagine resulted in the same escaped output; e.g.,
eval(new String("this is a \u201ctest&#amp;92;u201d"));
or even
eval(new String("this is a \u201ctest&#amp;92;u201d".replace('/&amp#92;/g','\')));
all results in the same thing:
"this is a \u201ctest\u201d";
It's as if I need to get the Javascript engine to somehow re-evaluate or re-parse the string, but I don't know what would do it. I thought perhaps eval() or just creating a new string from using the properly escaped input would do it, but now luck.
The fundamental question is - what do I have to do to turn the given string:
"this is a \u201ctest&#amp;92;u201d"
into a string that uses the proper Unicode characters?

Categories

Resources