Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Remove showModalDialog() #374

Merged
merged 1 commit into from
Dec 15, 2015
Merged

Remove showModalDialog() #374

merged 1 commit into from
Dec 15, 2015

Conversation

domenic
Copy link
Member

@domenic domenic commented Dec 2, 2015

Fixes https://www.w3.org/Bugs/Public/show_bug.cgi?id=26437.

A surprisingly clean cut. I was hoping there would be more stuff we could simplify by this removal, but it all seems contained within that section. I would love it if someone could prove me wrong.

@zcorpan
Copy link
Member

zcorpan commented Dec 3, 2015

Changes LGTM. None of the definitions are used elsewhere. The spec has the same pausing concept for both showModalDialog and alert, AFAICT.

Firefox hasn't removed it yet. https://bugzilla.mozilla.org/show_bug.cgi?id=981796
WebKit hasn't removed it yet. (I can't find any bug about removing it?)

Why is now a good time to remove it from the spec?

@domenic
Copy link
Member Author

domenic commented Dec 3, 2015

WebKit hasn't removed it yet. (I can't find any bug about removing it?)

WebKit has removed it from mobile, which is where the majority of WebKit usage is.

Why is now a good time to remove it from the spec?

Simply because I actually got around to making the pull request :). I think a feature that is in only 1.5 browsers and everyone agrees is deprecated should probably be removed, right?

@domenic domenic added the removal/deprecation Removing or deprecating a feature label Dec 3, 2015
@zcorpan
Copy link
Member

zcorpan commented Dec 3, 2015

It was never in mobile Safari to begin with, I believe.

Can you file a bug or something for WebKit?

@domenic
Copy link
Member Author

domenic commented Dec 4, 2015

WebKit bug filed https://bugs.webkit.org/show_bug.cgi?id=151885 /cc @hober @rniwa

@zcorpan
Copy link
Member

zcorpan commented Dec 10, 2015

Thanks. I suppose Mozilla is OK with removing showModalDialog from the spec, since @Ms2ger filed the original bug. Are WebKit people OK with it?

If nobody has objected by, I dunno, mid-January 2016, I suggest we merge this.

@Ms2ger
Copy link
Member

Ms2ger commented Dec 10, 2015

I do not represent either of Mozilla's browser teams on this (or anything, really).

@annevk
Copy link
Member

annevk commented Dec 14, 2015

LGTM. It's clear this is not going to be kept around much longer. Desktop Firefox and Safari don't have that power. (It does seem they might want to implement <dialog> first as some folks switched to that reading the various bug comments.)

@annevk
Copy link
Member

annevk commented Dec 14, 2015

(Somewhat sad that removing this doesn't allow us to remove nested event loops. What else is blocking that?)

@domenic
Copy link
Member Author

domenic commented Dec 14, 2015

I wasn't able to really find how "nested event loops" manifest in the spec.

@annevk
Copy link
Member

annevk commented Dec 14, 2015

Seems like that is "spin the event loop" which is used by a number of things (though not all of them manifest in nested event loops I suppose).

@annevk annevk merged commit eec9646 into master Dec 15, 2015
@annevk annevk deleted the rm-showmodaldialog branch December 15, 2015 16:38
@miketaylr
Copy link
Member

We want to remove showModalDialog from e10s Firefox, but have concerns that it's not web compatible (Outlook Web App and Exchange being the culprits we know about).

See https://bugzilla.mozilla.org/show_bug.cgi?id=1206885 (which is about attachments not working in Outlook Web App 2010). A quick google search will find you countless people complaining about OWA attachments not working in Edge or Chrome (with some people suggesting using Firefox as a workaround, or to install Silverlight :/):

http://answers.microsoft.com/en-us/windows/forum/apps_windows_10-msedge/outlook-web-access-owa-problems-in-edge/1de9d78d-dba0-46cb-9654-be7fd985b57f?page=5
https://code.google.com/p/chromium/issues/detail?id=421867
https://code.google.com/p/chromium/issues/detail?id=410283
https://productforums.google.com/forum/#!topic/chrome/F180I4IS77k
and more...

The bad news is that these OWA improvements are only available in Office 365 and there’s no sign of similar progress in either EAC or OWA for on-premises Exchange yet. Don’t hold your breath for an update in Exchange 2013 CU7 because I doubt that any change will happen there, even if its release was delayed last week.

source

@domenic I can't find the comment now, but I read in a Chromium bug that Outlook and Google were in communication about the issue, would you happen to know if there was a resolution?

@jacobrossi does Microsoft have a plan for this?

It feels wrong to remove this and break this many sites, is there something crazy we can do like shim showModalDialog in the spec (somehow)?

@jacobrossi
Copy link

We deprecated it a long while back. Edge does not have this API. It is still in IE and will remain there given lots of line of business apps designed for IE require it.

@miketaylr
Copy link
Member

We deprecated it a long while back. Edge does not have this API. It is still in IE and will remain there given lots of line of business apps designed for IE require it.

Right, I link to some forums above of people complaining about OWA 2010 not working in Edge. :/

@jacobrossi
Copy link

@miketaylr correct, which is why we recommend IE for legacy enterprise applications on Windows 10 and provide a way to jump from Edge to IE for those legacy apps. Historically, providing updates to old versions of OWA (pre-Office365 service based model) doesn't help in these situations because of on-prem installations that we don't manage. Moreover, on-prem installations typically don't show up in things like crawler statistics or UseCounters (IT Policies often disables this collection), so we also can't reliably determine when usage is sufficiently low. Hence, we freeze IE where it's at for these customers that still require legacy APIs while moving Edge ahead into the modern web without this cruft.

@zcorpan
Copy link
Member

zcorpan commented Jan 14, 2016

is there something crazy we can do like shim showModalDialog in the spec (somehow)?

I don't think so. Often the blocking behavior is needed, and you can't polyfill that without changing the client code, so filling this gap needs to happen in the apps that need it, I believe. Trying to polyfill it in the browser would probably cause more trouble.

@domenic
Copy link
Member Author

domenic commented Jan 14, 2016

We want to remove showModalDialog from e10s Firefox, but have concerns that it's not web compatible (Outlook Web App and Exchange being the culprits we know about).

I mean, depends on what you mean by web-compatible. At this point it's basically a Firefox-only feature (and Safari desktop). So it's not really part of the web, just the Firefox web.

@domenic I can't find the comment now, but I read in a Chromium bug that Outlook and Google were in communication about the issue, would you happen to know if there was a resolution?

@jeisinger might know more. My understanding is that the latest versions of OWA do not use showModalDialog.

@jacobrossi
Copy link

My understanding is that the latest versions of OWA do not use showModalDialog

Correct

@jeisinger
Copy link
Member

yes

@miketaylr
Copy link
Member

At this point it's basically a Firefox-only feature (and Safari desktop). So it's not really part of the web, just the Firefox web.

Well, Firefox (for now anyways) and Safari and IE. Just not Blink browsers or Edge. I'm not sure how that makes it not part of the web (even if it's really gross). It just feels weird to break the presumably thousands of OWA 2010 installations out there with the recommendation to stay on legacy IE browsers (or Safari, I guess) if you want to keep using your webmail.

@jeisinger
Copy link
Member

The recommendation is to upgrade to a supported version of exchange...

@miketaylr
Copy link
Member

The recommendation is to upgrade to a supported version of exchange...

Yes, that's the right recommendation for the site owners. Unfortunately software upgrades are more frequently business decisions than technical ones. I don't suspect anybody has teams doing evangelism for these sites (and my small team doesn't have the bandwidth).

(And generally the people in charge of these types of things aren't reading deprecation articles on Dev Opera or anywhere else.)

For the people commenting in Chromium bugs and in Microsoft forums (and in Bugzilla for e10s users), the recommendation is currently is "try another browser".

@annevk
Copy link
Member

annevk commented Jan 15, 2016

@miketaylr note that removal from the HTML standard does not mean that Firefox e10s cannot implement this. It just means that the authors of the HTML standard agreed that there's no future for this feature.

If Edge, Chrome, mobile Safari, Firefox e10s, all feel compelled to add this again, we'll obviously reconsider, but it seems far more likely that in a couple of years nobody has this anymore (except perhaps desktop Safari, their feature removal process seems extremely conservative).

@domenic domenic mentioned this pull request Jun 9, 2016
@yuhong
Copy link

yuhong commented Jun 9, 2016

FYI, Exchange 2010 is still supported and they released an update rollup to fix this long ago.

foolip added a commit to web-platform-tests/wpt that referenced this pull request Jul 9, 2018
These were added in commit 818bcab
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
foolip added a commit to web-platform-tests/wpt that referenced this pull request Jul 10, 2018
These were added in commit 818bcab
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
moz-v2v-gh pushed a commit to mozilla/gecko-dev that referenced this pull request Jul 20, 2018
…html/, a=testonly

Automatic update from web-platform-testsRemove .gitkeep files in 2dcontext/ and html/ (#11851)

These were added in commit 818bcab4ba58d2bc8dba94276c8b944c76570a19
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
--

wpt-commits: ddfe9c089bab565a9d3aa37bdef63d8012c1a94c
wpt-pr: 11851
jankeromnes pushed a commit to jankeromnes/gecko that referenced this pull request Jul 23, 2018
…html/, a=testonly

Automatic update from web-platform-testsRemove .gitkeep files in 2dcontext/ and html/ (#11851)

These were added in commit 818bcab4ba58d2bc8dba94276c8b944c76570a19
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
--

wpt-commits: ddfe9c089bab565a9d3aa37bdef63d8012c1a94c
wpt-pr: 11851
gecko-dev-updater pushed a commit to marco-c/gecko-dev-comments-removed that referenced this pull request Oct 3, 2019
…html/, a=testonly

Automatic update from web-platform-testsRemove .gitkeep files in 2dcontext/ and html/ (#11851)

These were added in commit 818bcab4ba58d2bc8dba94276c8b944c76570a19
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
--

wpt-commits: ddfe9c089bab565a9d3aa37bdef63d8012c1a94c
wpt-pr: 11851

UltraBlame original commit: 9a8d6a00920ec6fb7cb3dae6074ea4e26aaf5744
gecko-dev-updater pushed a commit to marco-c/gecko-dev-wordified-and-comments-removed that referenced this pull request Oct 3, 2019
…html/, a=testonly

Automatic update from web-platform-testsRemove .gitkeep files in 2dcontext/ and html/ (#11851)

These were added in commit 818bcab4ba58d2bc8dba94276c8b944c76570a19
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
--

wpt-commits: ddfe9c089bab565a9d3aa37bdef63d8012c1a94c
wpt-pr: 11851

UltraBlame original commit: 9a8d6a00920ec6fb7cb3dae6074ea4e26aaf5744
gecko-dev-updater pushed a commit to marco-c/gecko-dev-wordified that referenced this pull request Oct 3, 2019
…html/, a=testonly

Automatic update from web-platform-testsRemove .gitkeep files in 2dcontext/ and html/ (#11851)

These were added in commit 818bcab4ba58d2bc8dba94276c8b944c76570a19
and reflect the structure of the HTML spec as it was in 2012.

The structure largely remains the same, but some directories like
webappapis/user-prompts/dialogs-implemented-using-separate-documents/
are gone from the spec (whatwg/html#374) and
others like introduction/ are non-normative and should never be
populated.

But mainly, a lot of empty directories is a nuisance when exploring
the directory in a file manager or on GitHub.

The number of directories in 2dcontext/ and html/ combined decreases
by almost 200, from 653 to 456.
--

wpt-commits: ddfe9c089bab565a9d3aa37bdef63d8012c1a94c
wpt-pr: 11851

UltraBlame original commit: 9a8d6a00920ec6fb7cb3dae6074ea4e26aaf5744
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
removal/deprecation Removing or deprecating a feature
Development

Successfully merging this pull request may close these issues.

8 participants