Disable browser password save: revisited

[Alexander Bulei]

Hi magosk,

Teamviewer will be fine and enough

Best Regards.

[magosk]

If it is possible for you, we can schedule a TeamViewer session today. You can contact me by replying to this post or contacting me on Skype: magnus.oskarsson-enera.se.

/Magnus

[Alexander Bulei]

Hi magosk,

Contact us by email, to schedule the remote session.

Best Regards.

[magosk]

Hi!

As you may recall, we never solved installing the newer CGDevTools versions with IW 14.0.63 (XE7). Finally, we have now received a new IW version (14.2.2) that works well for us and includes some requested security features. Now I have been able to successfully install the newer CGDevTools (I currently have 4.0.1.79 on my dev. computer), and we can install our custom package as well. So far so good.

But I have tested the original problem again, and it unfortunately remains. To repeat it: we have a login page that has both a (company) subscription ID and password field, as well as a user name and password (within the company subscription) field. Also, there is a hidden passcode field which is used under some circumstances (two-factor login). Despite all TIWCGJQMEdit controls on the page having JQMEditOptions.EditTypeOptions.Autocomplete = False, Chrome Mobile still asks if you want to save the user credentials. And if you ask Yes and go to the login page again, Chrome has problems with multiple password fields and fills in the same value for all of them (e.g. if you have a blank user password, Chrome falsely sets the subscription password in this field, and in addition in the hidden passcode field, which totally screws up our login logic). I know you tried to make a fix for this earlier, but either it is not working, or I am missing something I need to do in my code.

If you want I can give you a URL and some login credentials if you want to see the behaviour yourself (and perhaps analyze the HTML/JS code).

Best regards

Magnus Oskarsson

[Jorge Sousa]

Hi

About

I know you tried to make a fix for this earlier, but either it is not working, or I am missing something I need to do in my code.

Honestly, its the first time i read something about it.

If you want I can give you a URL and some login credentials if you want to see the behaviour yourself (and perhaps analyze the HTML/JS code).

Please do, this must be related to one of these matters:

https://stackoverflow.com/questions/35863928/how-to-hint-google-smart-lock-to-use-a-specific-input-field-as-the-username-in-a
https://stackoverflow.com/questions/23156578/google-chrome-autofilling-all-password-inputs
https://stackoverflow.com/questions/15431783/html-chrome-save-password

[magosk]

Hi Jorge and thanks for the quick reply!

If you look at the beginning of this thread, Alexander says he made an implementation which was part of the 3.5.0.218 development build. But he refers to duplicating an old change for the jQuery UI edit, and since Chrome Mobile did not behave like this earlier it may be that something else is required now. Anyway, I will send URL and login details to your support e-mail. It will be an empty account, so you won't be able to do much except login and logout, but that is all what is needed in this case

You will note that we have our own support for storing login details (on the server side), but a security-oriented customer may disable storing of subscription and/or user credentials if so desired. That is why we do not want browsers to store user credentials on the client side. Please also note that you can go into Settings for the Chrome Mobile browser and delete any previous choices you made for this web-site regarding saving the password or not. And as mentioned, you get the problem the next time you attempt to login after answering Yes.

/Magnus

[Jorge Sousa]

Hi

Ok, send us the demo's url here or by email

Or even better, send us a test case demo

[magosk]

Hi,

I sent an e-mail to your support address just after the last post. Perhaps it was caught as spam due to the url link?

[Jorge Sousa]

Hi

We received the email

so in what platform should we test first for chrome ?

[Alexander Bulei]

Hi magosk,

It's a browser feature, which detect if the form contain password field, will open the dialog to save the password..

As workarround, for your system, you try to generate the random suffix or prefix for edit name:

delphi code

procedure TIWMainForm.IWAppFormCreate(Sender: TObject);
begin
  IWCGJQEdit1.Name:= 'EDIT_'+IntToStr(DateTimeToUnix(Now))+'_PW';
end;

The browser, will still asking to save, even if you answer "Yes", on next load, will NOT fill the inputs, since they have another "ID" and "NAME" attributes.

Best Regards.