UNCONFIRMED 28878
MSP_Refresh XMLHttpRequest possible onreadystatechange duplicate calls
https://bugs.webkit.org/show_bug.cgi?id=28878
Summary XMLHttpRequest possible onreadystatechange duplicate calls
John Grange
Reported 2009-09-01 09:46:14 PDT
The product "msp center plus" by adventnet produces a bug in Safari 2.x - 4.x and Latest webkit nightly builds. When browsing the site (intranet site not there home page) clicking on certain items will result in a continuous refresh of the page between the previous and new page. This is not present in other browsers (Firefox / IE). Please let me know how I can help to reproduce the bug for you since the product page is internal? I am "guessing" this is a javascript problem but I could be wrong.
Attachments
John Grange
Comment 1 2009-09-01 10:04:03 PDT
There is a demo page on the above URL but it appears to be offline at this time.
John Grange
Comment 2 2009-09-05 10:49:43 PDT
The demo page is back online, testing needs to be done mainly in the network administrator section (click network monitoring on upper left side).
Daniel Bates
Comment 3 2009-09-07 17:31:25 PDT
I am unable to reproduce the bug. Are there specific links on the page that trigger it. (In reply to comment #2) > The demo page is back online, testing needs to be done mainly in the network > administrator section (click network monitoring on upper left side).
Daniel Bates
Comment 4 2009-09-07 17:35:15 PDT
What revision of WebKit are you using? You can find this out by selecting About Safari from the Safari menu. The revision number should be of the form r480xx.
John Grange
Comment 5 2009-09-07 18:35:51 PDT
I just tested with the latest version (Version 4.0.3 (6531.9, r48139)), and the bug is still present. It is hard to describe how to trigger it. I have included an example url that it triggered on. https://demo.mspcenterplus.com/RIMHome.do?mspType=Monitoring#/devices/objectdetails.do?requestid=SNAPSHOT&name=192.168.105.187.opmgwin2 The best way to trigger the bug is to click in networking monitoring and then just start clicking on random items , it is not always the same link and sometimes takes awhile before it triggers, but it is always happens in the network monitoring.
Daniel Bates
Comment 6 2009-09-14 21:35:12 PDT
I was able to trigger the issue. It is a JavaScript issue. However, I haven't yet confirmed whether this is a bug in WebKit/JavaScriptCore. (In reply to comment #5) > I just tested with the latest version (Version 4.0.3 (6531.9, r48139)), and the > bug is still present. It is hard to describe how to trigger it. I have included > an example url that it triggered on. > > https://demo.mspcenterplus.com/RIMHome.do?mspType=Monitoring#/devices/objectdetails.do?requestid=SNAPSHOT&name=192.168.105.187.opmgwin2 > > The best way to trigger the bug is to click in networking monitoring and then > just start clicking on random items , it is not always the same link and > sometimes takes awhile before it triggers, but it is always happens in the > network monitoring.
John Grange
Comment 7 2009-09-19 08:38:27 PDT
Is there anything more I can do on my end?
Daniel Bates
Comment 8 2009-09-19 14:25:34 PDT
Yes. If you want to try to expedite this and have some time, please provide a reduction. That is, come up with a concise test case that reproduces this issue. Last I recall, I was focusing around the sendAjaxRequest (declared in file /webclient/common/js/compressed1.js) call, and the external JavaScript /webclient/common/js/rsh.js (i.e the dhtmlHistory component <http://code.google.com/p/reallysimplehistory/>). From my understanding, sendAjaxRequest sets the onreadystatechange handler to call processResponseHTML (also declared in compressed1.js), which calls dhtmlHistory.add(), which builds a sack of the visited URLs (i.e. a history). If I recall correctly, I noticed that when the issue occurs the same two URLs would get continuously added to this stack. I believe the dhtmlHistory.add() is only call by the onreadystatechange handler, which made me skeptical of whether it was firing duplicate messages. (In reply to comment #7) > Is there anything more I can do on my end?
Note You need to log in before you can comment on or make changes to this bug.