Source code
Revision control
Copy as Markdown
Other Tools
/* -*- indent-tabs-mode: nil; js-indent-level: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
/*
*
* Date: 09 November 2002
* SUMMARY: Test that interpreter can handle string literals exceeding 64K
*
* Test that the interpreter can handle string literals exceeding 64K limit.
* For that the script passes to eval() "str ='LONG_STRING_LITERAL';" where
* LONG_STRING_LITERAL is a string with 200K chars.
*
* Igor Bukanov explains the technique used below:
*
* > Philip Schwartau wrote:
* >...
* > Here is the heart of the testcase:
* >
* > // Generate 200K long string
* > var long_str = duplicate(LONG_STR_SEED, N);
* > var str = "";
* > eval("str='".concat(long_str, "';"));
* > var test_is_ok = (str.length == LONG_STR_SEED.length * N);
* >
* >
* > The testcase creates two identical strings, |long_str| and |str|. It
* > uses eval() simply to assign the value of |long_str| to |str|. Why is
* > it necessary to have the variable |str|, then? Why not just create
* > |long_str| and test it? Wouldn't this be enough:
* >
* > // Generate 200K long string
* > var long_str = duplicate(LONG_STR_SEED, N);
* > var test_is_ok = (long_str.length == LONG_STR_SEED.length * N);
* >
* > Or do we specifically need to test eval() to exercise the interpreter?
*
* The reason for eval is to test string literals like in 'a string literal
* with 100 000 characters...', Rhino deals fine with strings generated at
* run time where lengths > 64K. Without eval it would be necessary to have
* a test file excedding 64K which is not that polite for CVS and then a
* special treatment for the compiled mode in Rhino should be added.
*
*
* >
* > If so, is it important to use the concat() method in the assignment, as
* > you have done: |eval("str='".concat(long_str, "';"))|, or can we simply
* > do |eval("str = long_str;")| ?
*
* The concat is a replacement for eval("str='"+long_str+"';"), but as
* long_str is huge, this leads to constructing first a new string via
* "str='"+long_str and then another one via ("str='"+long_str) + "';"
* which takes time under JDK 1.1 on a something like StrongArm 200MHz.
* Calling concat makes less copies, that is why it is used in the
* duplicate function and this is faster then doing recursion like in the
* test case to test that 64K different string literals can be handled.
*
*/
//-----------------------------------------------------------------------------
var UBound = 0;
var BUGNUMBER = 179068;
var summary = 'Test that interpreter can handle string literals exceeding 64K';
var status = '';
var statusitems = [];
var actual = '';
var actualvalues = [];
var expect= '';
var expectedvalues = [];
var LONG_STR_SEED = "0123456789";
var N = 20 * 1024;
var str = "";
// Generate 200K long string and assign it to |str| via eval()
var long_str = duplicate(LONG_STR_SEED, N);
eval("str='".concat(long_str, "';"));
status = inSection(1);
actual = str.length == LONG_STR_SEED.length * N
expect = true;
addThis();
//-----------------------------------------------------------------------------
test();
//-----------------------------------------------------------------------------
function duplicate(str, count)
{
var tmp = new Array(count);
while (count != 0)
tmp[--count] = str;
return String.prototype.concat.apply("", tmp);
}
function addThis()
{
statusitems[UBound] = status;
actualvalues[UBound] = actual;
expectedvalues[UBound] = expect;
UBound++;
}
function test()
{
printBugNumber(BUGNUMBER);
printStatus(summary);
for (var i=0; i<UBound; i++)
{
reportCompare(expectedvalues[i], actualvalues[i], statusitems[i]);
}
}