Labelling with current canary milestone for triaging.

When I go to the URL you mentioned, I get a login screen. What are the credentials?

The given URl navigates to the login page . 

@oneoftheevil-- Please update the thread as per comment#2 and also provide us the chrome version and OS details , in which you faced the issue.

Thanks!

Sorry,
with this link you should be logged in 
http://itsm-demo.otrs.com/otrs/index.pl?Action=Login&User=changemanager-E&Password=otrs123
Best
Dave

Notes from my debugging session -
- The issue is that the submit button is not included as a field in the POST request body.
- This probably happens because when you click on "Submit", you click on elements within the <button>, which breaks some code that adds <input type=hidden name=foo> where "foo" is the name of the target element of the submission button click, which is the <span> within it instead of <button, so the name is empty.
- Firefox does not dispatch "click" to elements within <button>. See https://bugzilla.mozilla.org/show_bug.cgi?id=1089326 (this may change at some point and break you as well).
- If you click on the border of the "Submit" button, the window closes.

In case this is a new issue that only started since Chrome 54, it might be related to  issue 196640 , however, I am still investigating whether this is the case.

More notes -
- I tried to remove that <input type="hidden" name=foo> addition and the bug still occurs.
- I tried to remove the contents of the <button> and the bug does not occur anymore.

The form is not submitted using the normal HTML form submission form (despite being triggered by a <button type=submit> click) - there is some validation (jQuery validation) process and then the form is submitted using form.submit() and the "submit" event returns false (prevents the default action, does not submit again).

I looked at the HTML specification and it seems that the behavior compliant.
It says (in my own words) -
- When using the form.submit() method, the submitter element is the form itself - https://html.spec.whatwg.org/#the-form-element:concept-form-submit
- Only if a <button> is the submitter, include it in the form-data (the data that is sent to the server upon submission) - https://html.spec.whatwg.org/#constructing-form-data-set

Since form.submit() is used, the submitter is always the form, so <button> should never be included in the form-data.

So this is basically a jQuery Validation issue, because it listens to the "click" event on elements the ":submit" jQuery selector (which selects <input type=submit> and <button type=submit>) selects.
I submitted a pull request to jQuery Validation - https://github.com/jzaefferer/jquery-validation/pull/1901, but OTRS must update its plugin version (once it is released) in order to get the fix, or use a CSS workaround. Its GitHub organization disables the issue tracker, so I cannot file an issue.

The CSS workaround would be -
button[type="submit"] *
{
 pointer-events: none;
}

A more localized and more performant CSS workaround would be -
button[type="submit"] span
{
 pointer-events: none;
}


By the way, pressing Enter on the various field (it is supposed to generate a synthetic "click" event on the submit button) does not fix the problem because there are numerous submit buttons and nothing makes the last one (the "Submit" button) get a synthetic "click" event.


To summarize, I think this is a WontFix.
Was this always a problem? Did it work in earlier versions of Chrome? If so, which version (or a period of time) did not have the issue that you can remember?

The issue can be reproduced, so it is not uncomfirmed, however, it might not be a browser issue, but coding issue.

rbyers@, tkent@ - seems related to issue  crbug.com/519621  - what do you think? WontFix?

Yeah, I think this is a dupe of  Issue 519621 .  phistuck@, thank you for the investigation.

Note - my pull request was accepted by the jQuery Validation team (for now, anyway) -
https://github.com/jquery-validation/jquery-validation/pull/1901
If you update jQuery Validation in OTRS (I do not know about its release frequency, though) to one that has my fix, it should resolve the issue.