121931 - ActiveX control - prompt changes break form submit

Status: RESOLVED FIXED

(Core Graveyard :: Embedding: ActiveX Wrapper, defect)

Description

[Aaron Graves]

A "submit" style button that submits information entered into a web from does
not work in the ActiveX control. Information can be entered, but when the
"submit" button is clicked, the browser does nothing.

Comment 1

[Damian Yerrick]

Reporter: In order to fix this bug, we need a bit more information from you.

  Steps to reproduce (including URL):

  Actual results:

  Expected results:

I doubt that Mozilla even claims to support ActiveX controls.

Comment 2

[Aaron Graves]

This isn't about supporting ActiveX controls... I'm talking about the Mozilla
ActiveX Control. When I use it in my program, and when you go to any webpage
that accepts text input and has some kind of a submit button (try any web search
site, such as www.altavista.com), when you click the submit button it does
nothing, whereas it should submit the data in the textbox.

Comment 3

[Adam Lock]

Confirming.

It looks as though the form submit warning about submitting data on an insecure
channel is being displayed via nsIPrompt::ConfirmEx. The ActiveX controls prompt
service hasn't implemented ConfirmEx so nothing is displayed. It used to be
displayed with ConfirmCheck so it worked fine before now.

The workaround for the time being is add a line like this to MozillaControl
profile's prefs.js so the prompt is not displayed when submitting forms.

user_pref("security.warn_submit_insecure", false);

Comment 4

[Adam Lock]

Attachment: Patch fixes the problem

This patch fixes the problem.

It's pretty large because I bit the bullet and wrote a dialog class that
handles ConfirmEx, ConfirmCheck, Prompt, PromptPassword &
PromptUsernameAndPassword. It can be extended at some point to also support
AlertCheck & Select but this will suffice for now.

Comment 5

[Adam Lock]

Chak, could you review this for me please? Thanks

Comment 6

[Chak Nanga]

Comment on attachment 66594 [details] [diff] [review]
Patch fixes the problem

r=chak

Comment 7

[Asa Dotzler [:asa]]

a=asa (on behalf of drivers) for checkin to 0.9.8

Comment 8

[Adam Lock]

Fix is checked in 0.9.8 branch and trunk