Bug 4372

Summary:

JavaScript document.write page form does not submit in Safari

Product:

WebKit

Reporter:

Vicki Murley <vicki>

Component:

DOM

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

ap, gavin.sharp, ian, mitz

Priority:

Keywords:

HasReduction, InRadar

Version:

420+

Hardware:

Mac

OS:

OS X 10.4

Attachments:

Description	Flags
test case	none
patch, without tests or ChangeLog	mjs: review-
proposed fix	darin: review+

Vicki Murley

Reported 2005-08-10 11:23:04 PDT

This bug is also in Radar as <rdar://4196379> Summary: Submitting an html form of a page that was written with a JavaScript document.write method does not work in Safari. We are concerned about this problem because we are using an encryption/decryption product that has the ability to decrypt a page and uses a document.write method to display the page. Steps to Reproduce: I have created a small page that demonstrates the problem. On load of the page it will use a document.write method to over write the current page with a page that has a form on it. Upon clicking on the submit button Safari does not submit the form; however Internet Explorer 6(WinXP) and Firefox (Mac) submits the form correctly. Example page writeform.html: <html> <head> <script> function writePage() { document.write("<html><head></head><body>" + "<form name='aform' action='servlet.html' method='post'>" + "<input type=submit></form>" + "</body></html>"); } </script> </head> <body onload="writePage()"> </body> </html> Expected Results: Safari would submit the form when clicking on the submit button. Actual Results: Safari does not submit the form and stays on current page. Regression: This problem occurred on the following platforms: Mac os X 10.3.9 Safari 1.3(v312) PowerBook G4 15-inch A1095. Mac os X 10.4.2 Safari 2.0(412.3) PowerBook G4 15-inch A1095. -------------------------------------------

Attachments
test case (279 bytes, text/html) 2005-08-10 11:23 PDT, Vicki Murley	no flags	Details
patch, without tests or ChangeLog (531 bytes, patch) 2006-09-03 10:37 PDT, Alexey Proskuryakov	mjs: review-	Details Formatted Diff Diff
proposed fix (3.60 KB, patch) 2006-09-23 20:34 PDT, Alexey Proskuryakov	darin: review+	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Vicki Murley

Comment 1 2005-08-10 11:23:50 PDT

Created attachment 3313 [details] test case

Eric Seidel (no email)

Comment 2 2005-12-27 15:00:00 PST

Odd. When I hit the back button (after having hit submit), it tries to take me to /servlet.html

Alexey Proskuryakov

Comment 3 2006-09-03 10:37:50 PDT

Created attachment 10379 [details] patch, without tests or ChangeLog All layout tests pass, but it might be just because the reason for the removed check isn't covered... So, I'm asking about validity of such approach. A workaround is to call document.close() after writing.

Darin Adler

Comment 4 2006-09-03 12:11:19 PDT

Comment on attachment 10379 [details] patch, without tests or ChangeLog Oof! We *definitely* need tests for this one. I can't tell just from code inspection if this is OK or not. I'm not sure what to do about reviewing.

Alexey Proskuryakov

Comment 5 2006-09-14 08:38:35 PDT

The code in question appeared in WebCore in r1035: * src/kdelibs/khtml/khtml_part.cpp: Added. Not used yet, but maybe we'll use part of it some time soon. In the KDE repository, the condition has been tweaked many times, but the part I'm removing was introduced in r64718 six years ago: "applying stripped down version of my anti-tokenizer-crash-on-submit patch discussed on kfm-devel. It only prevents the crash. A complete solution for the form data has still to be found." http://websvn.kde.org/trunk/kdelibs/khtml/khtml_part.cpp?rev=64718&r1=64461&r2=64718 I couldn't find the mentioned kfm-devel discussion.

Maciej Stachowiak

Comment 6 2006-09-23 01:18:00 PDT

I think the approach is right but this needs a layout test.

Alexey Proskuryakov

Comment 7 2006-09-23 20:34:50 PDT

Created attachment 10730 [details] proposed fix

mitz

Comment 8 2006-09-27 14:42:30 PDT

(In reply to comment #5) > I couldn't find the mentioned kfm-devel discussion. http://lists.kde.org/?l=kfm-devel&m=96644872718833&w=2

Darin Adler

Comment 9 2006-09-28 09:09:22 PDT

Comment on attachment 10730 [details] proposed fix r=me

Alexey Proskuryakov

Comment 10 2006-10-01 01:41:09 PDT

Committed revision 16694. The fun begins!

Note You need to log in before you can comment on or make changes to this bug.