Attachment #118281 for bug #72762

View Details Raw Unified Return to Bug 72762

| Differences between



<html>
<head>
<link id="img" rel="stylesheet" href="resources/stylesheet304.php">
</head>
<body>
It's possible for a successful revalidation to cause events to fire synchronously.
We shouldn't crash if one of those events stops the load.
See https://bugs.webkit.org/show_bug.cgi?id=72762.
<script>
if (window.layoutTestController) {
    layoutTestController.dumpAsText();
    layoutTestController.waitUntilDone();
}

function obl(e) {
    if (window.sessionStorage.getItem("reloaded")) {
        window.stop();
        window.sessionStorage.removeItem("reloaded");
        setTimeout(finish, 0);
    } else {
        window.sessionStorage.reloaded = "true";
        location.reload();
    }
}

function finish() {
    if (window.layoutTestController)
        layoutTestController.notifyDone();
}
</script>
<object type="application/x-no-such-mime" onbeforeload="obl();"></embed>
</body>
</html>

Line 0 LayoutTests/http/tests/cache/cancel-during-revalidation-succeeded-expected.txt_sec1
- LayoutTests/http/tests/cache/cancel-during-revalidation-succeeded-expected.txt +1 lines
		1	It's possible for a successful revalidation to cause events to fire synchronously. We shouldn't crash if one of those events stops the load. See https://bugs.webkit.org/show_bug.cgi?id=72762.



<?php
require_once '../../resources/portabilityLayer.php';

clearstatcache();

if ($_SERVER["HTTP_IF_MODIFIED_SINCE"]) {
    header("HTTP/1.0 304 Not Modified");
    exit();
}
$one_year = 12 * 31 * 24 * 60 * 60;
$last_modified = gmdate(DATE_RFC1123, time() - $one_year);
$expires = gmdate(DATE_RFC1123, time() + $one_year);

header('Cache-Control: public, max-age=' . $one_year);
header('Expires: ' . $expires);
header('Content-Type: text/css');
header('Content-Length: 0');
header('Etag: 123456789');
header('Last-Modified: ' . $last_modified);
exit();
?>



2011-12-07  Nate Chapin  <japhet@chromium.org>

        Test for https://bugs.webkit.org/show_bug.cgi?id=72762

        Reviewed by NOBODY (OOPS!).

        * http/tests/cache/cancel-during-revalidation-succeeded-expected.txt: Added.
        * http/tests/cache/cancel-during-revalidation-succeeded.html: Added.
        * http/tests/cache/resources/stylesheet304.php: Added.

2011-12-07  James Simonsen  <simonjam@chromium.org>

        [Chromium] Unreviewed, clean out passing tests from test_expectations.txt



2011-12-07  Nate Chapin  <japhet@chromium.org>

        A SubresourceLoader in the middle of revalidating
        a resource should be treated as finishing (similar to
        didFinishLoading and didFail) to ensure that willCancel()
        doesn't declare the revalidation as having failed reentrantly.
        https://bugs.webkit.org/show_bug.cgi?id=72762

        Reviewed by NOBODY (OOPS!).

        Test: http/tests/cache/cancel-during-revalidation-succeeded.html

        * loader/SubresourceLoader.cpp:
        (WebCore::SubresourceLoader::didReceiveResponse):

2011-12-07  Andreas Kling  <kling@webkit.org>

        RenderObject::style(): Inline early-return condition.

Lines 148-155 Source/WebCore/loader/SubresourceLoader.cpp_sec1
- Source/WebCore/loader/SubresourceLoader.cpp -1 / +5 lines
148	if (response.httpStatusCode() == 304) {	148	if (response.httpStatusCode() == 304) {
149	// 304 Not modified / Use local copy	149	// 304 Not modified / Use local copy
150	// Existing resource is ok, just use it updating the expiration time.	150	// Existing resource is ok, just use it updating the expiration time.
		151	m_state = Finishing;
151	memoryCache()->revalidationSucceeded(m_resource, response);	152	memoryCache()->revalidationSucceeded(m_resource, response);
152	ResourceLoader::didReceiveResponse(response);	153	if (!reachedTerminalState()) {
		154	ResourceLoader::didReceiveResponse(response);
		155	cancel();
		156	}
153	return;	157	return;
154	}	158	}
155	// Did not get 304 response, continue as a regular resource load.	159	// Did not get 304 response, continue as a regular resource load.

Return to Bug 72762

Line 0 LayoutTests/http/tests/cache/cancel-during-revalidation-succeeded.html_sec1
- LayoutTests/http/tests/cache/cancel-during-revalidation-succeeded.html +33 lines
	1	<html>
	2	<head>
	3	<link id="img" rel="stylesheet" href="resources/stylesheet304.php">
	4	</head>
	5	<body>
	6	It's possible for a successful revalidation to cause events to fire synchronously.
	7	We shouldn't crash if one of those events stops the load.
	8	See https://bugs.webkit.org/show_bug.cgi?id=72762.
	9	<script>
	10	if (window.layoutTestController) {
	11	layoutTestController.dumpAsText();
	12	layoutTestController.waitUntilDone();
	13	}
	14
	15	function obl(e) {
	16	if (window.sessionStorage.getItem("reloaded")) {
	17	window.stop();
	18	window.sessionStorage.removeItem("reloaded");
	19	setTimeout(finish, 0);
	20	} else {
	21	window.sessionStorage.reloaded = "true";
	22	location.reload();
	23	}
	24	}
	25
	26	function finish() {
	27	if (window.layoutTestController)
	28	layoutTestController.notifyDone();
	29	}
	30	</script>
	31	<object type="application/x-no-such-mime" onbeforeload="obl();"></embed>
	32	</body>
	33	</html>

Line 0 LayoutTests/http/tests/cache/resources/stylesheet304.php_sec1
- LayoutTests/http/tests/cache/resources/stylesheet304.php +21 lines
	1	<?php
	2	require_once '../../resources/portabilityLayer.php';
	3
	4	clearstatcache();
	5
	6	if ($_SERVER["HTTP_IF_MODIFIED_SINCE"]) {
	7	header("HTTP/1.0 304 Not Modified");
	8	exit();
	9	}
	10	$one_year = 12 * 31 * 24 * 60 * 60;
	11	$last_modified = gmdate(DATE_RFC1123, time() - $one_year);
	12	$expires = gmdate(DATE_RFC1123, time() + $one_year);
	13
	14	header('Cache-Control: public, max-age=' . $one_year);
	15	header('Expires: ' . $expires);
	16	header('Content-Type: text/css');
	17	header('Content-Length: 0');
	18	header('Etag: 123456789');
	19	header('Last-Modified: ' . $last_modified);
	20	exit();
	21	?>

Lines 1-3 LayoutTests/ChangeLog_sec1
- LayoutTests/ChangeLog +10 lines
		1	2011-12-07 Nate Chapin <japhet@chromium.org>
		2
		3	Test for https://bugs.webkit.org/show_bug.cgi?id=72762
		4
		5	Reviewed by NOBODY (OOPS!).
		6
		7	* http/tests/cache/cancel-during-revalidation-succeeded-expected.txt: Added.
		8	* http/tests/cache/cancel-during-revalidation-succeeded.html: Added.
		9	* http/tests/cache/resources/stylesheet304.php: Added.
		10
1	2011-12-07 James Simonsen <simonjam@chromium.org>	11	2011-12-07 James Simonsen <simonjam@chromium.org>
2		12
3	[Chromium] Unreviewed, clean out passing tests from test_expectations.txt	13	[Chromium] Unreviewed, clean out passing tests from test_expectations.txt

Lines 1-3 Source/WebCore/ChangeLog_sec1
- Source/WebCore/ChangeLog +15 lines
		1	2011-12-07 Nate Chapin <japhet@chromium.org>
		2
		3	A SubresourceLoader in the middle of revalidating
		4	a resource should be treated as finishing (similar to
		5	didFinishLoading and didFail) to ensure that willCancel()
		6	doesn't declare the revalidation as having failed reentrantly.
		7	https://bugs.webkit.org/show_bug.cgi?id=72762
		8
		9	Reviewed by NOBODY (OOPS!).
		10
		11	Test: http/tests/cache/cancel-during-revalidation-succeeded.html
		12
		13	* loader/SubresourceLoader.cpp:
		14	(WebCore::SubresourceLoader::didReceiveResponse):
		15
1	2011-12-07 Andreas Kling <kling@webkit.org>	16	2011-12-07 Andreas Kling <kling@webkit.org>
2		17
3	RenderObject::style(): Inline early-return condition.	18	RenderObject::style(): Inline early-return condition.