It doesn't seem like the timing was part of that bug. Initially, when I was
involved at least, the timing was used to set mDialogAbuseCount so that we
could track if there was an abuse at all. Once the abuse count was hit, we
prompted the user and, if the user so chose, disabled dialogs for the
duration of that nsGlobalWindow (top window, otherwise frames could be an
easy workaround). That was my intent as I recall, but things may have
changed since I was involved (either in the followup patches by jst or in
follow up bugs).
Natch.
On Wed, Sep 18, 2013 at 7:56 AM, Gavin Sharp <ga...@gavinsharp.com> wrote:
> I remember discovering some of this confusing behavior while working
> on bug 391834. It dates back to bug 61098, and following the reasoning
> from those bug comments is a bit tricky. jst or Natch might recall the
> details, but I doubt it :)
>
> I wouldn't really assume that there's some great reason for the
> current behavior. I'm inclined to agree with you that we should just
> simplify the code and the behavior.
>
> Gavin
>
> On Tue, Sep 17, 2013 at 6:06 PM, Tim Abraldes <tabral...@mozilla.com>
> wrote:
> > This is a re-post from firefox-dev [3], in case anyone in dev-platform
> is aware of historical context. Please follow-up here in dev.platform (I
> think it's easier for users/mail-clients to jump into a newsgroup
> discussion than a mailing list discussion)
> >
> > The test page at [1] illustrates the current behavior of the "Prevent
> this page from creating additional dialogs" checkbox. Here also is a brief
> description:
> > 1) If a dialog is shown and fewer than
> "dom.successive_dialog_time_limit" milliseconds have elapsed since the last
> dialog was closed, the "Prevent this page from creating additional dialogs"
> checkbox is added to the dialog
> > 2) If the user does not select the checkbox, go to 1
> > 3) If the user does select the checkbox, then future dialogs on the
> page are limited to a rate of 1 dialog per
> "dom.successive_dialog_time_limit" milliseconds (e.g. two consecutive
> alert() calls may result in {failure, success} if the first call occurs
> 2999ms after the last dialog closed, and the second call occurs 1ms later.
> In this case, "failure" means that an exception is thrown by alert())
> >
> > This behavior is confusing to me - so confusing that I filed bug 910501
> [2] without realizing that this is intended behavior. As a user, I expect
> that a checkbox labeled "Prevent this page from creating additional
> dialogs" will prevent the page from creating ANY additional dialogs, and so
> I would (and did) file a bug when that expectation isn't met. I think this
> is not just a wording issue: I much prefer the ability to disable all
> dialogs on a page than the ability to rate-limit dialogs on that page.
> >
> > Of course, I don't speak for everyone, and I'm sure that there are
> reasons for the current behavior. I propose that we change the behavior of
> the "Prevent this page from creating additional dialogs" to prevent ALL
> subsequent dialogs on the page, but I'm interested in hearing
> ideas/arguments supporting the current behavior.
> >
> > [1] https://bug856977.bugzilla.mozilla.org/attachment.cgi?id=806066
> > [2] https://bugzilla.mozilla.org/show_bug.cgi?id=910501
> > [3]
> https://mail.mozilla.org/pipermail/firefox-dev/2013-September/000955.html
> > _______________________________________________
> > dev-platform mailing list
> > dev-platform@lists.mozilla.org
> > https://lists.mozilla.org/listinfo/dev-platform
>
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform
