893004 - Remove Navigator classinfo and most of its XPCOM goop

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Boris Zbarsky [:bzbarsky]]

Now that bug 838146 is fixed, we can do that.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: part 1. Remove Navigator classinfo.

Comment 2

[Boris Zbarsky [:bzbarsky]]

Attachment: part 2. Remove various no-longer-used XPCOM bits from Navigator, except the ones from nsIDOMNavigator.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: part 3. Remove the unused-from-C++ bits of nsIDOMNavigator, except the ones that share implementation with WebIDL anyway. that I believe we could remove all these remaining attributes too; it's just some tedious work.

Comment 4

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Comment on attachment 777099 [details] [diff] [review]
part 1.  Remove Navigator classinfo.

Review of attachment 777099 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/base/nsDOMClassInfo.cpp
@@ -4430,5 @@
> -      // don't need to worry about losing expando properties etc.
> -      if (!::JS_DefinePropertyById(cx, obj, id, v,
> -                                   JS_PropertyStub, JS_StrictPropertyStub,
> -                                   JSPROP_READONLY | JSPROP_PERMANENT |
> -                                   JSPROP_ENUMERATE)) {

Seems like it's unforgeable-ish right now; do we want to change that?

Comment 5

[Boris Zbarsky [:bzbarsky]]

Well, we already changed it, since the code you quote is dead code once we're on WebIDL bindings.

I don't think there's any good reason for window.navigator to be unforgeable.

Comment 6

[Boris Zbarsky [:bzbarsky]]

Er, I'm wrong.  It's not dead code.

I still think there's no good reason to have it be unforgeable.

Comment 7

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 777099 [details] [diff] [review]
part 1.  Remove Navigator classinfo.

Ok, based on Bug 342490 I think this should be ok.

Comment 8

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 777100 [details] [diff] [review]
part 2.  Remove various no-longer-used XPCOM bits from Navigator, except the ones from nsIDOMNavigator.

Could you fix the naming of onsucces

Rename nsIDOMNavigatorNetwork.idl to nsIMozNavigatorNetwork.idl

Comment 9

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 777101 [details] [diff] [review]
part 3.  Remove the unused-from-C++ bits of nsIDOMNavigator, except the ones that share implementation with WebIDL anyway.   that I believe we could remove all these remaining attributes too; it's just some tedious work.

This is a bit risky, but let's try.
Hopefully binary addons devs complain early enough.

Do we have the comment about vibrate() elsewhere? Same with requestWakeLock

Comment 10

[Boris Zbarsky [:bzbarsky]]

> Do we have the comment about vibrate() elsewhere? 

No, because the behavior it describes is no longer what happens, since WebIDL supersedes it.

> Same with requestWakeLock

In Navigator.webidl.

Comment 11

[Boris Zbarsky [:bzbarsky]]

Addressed the comments and:

https://hg.mozilla.org/integration/mozilla-inbound/rev/c80617d27cd9
https://hg.mozilla.org/integration/mozilla-inbound/rev/18233d9d3ed5
https://hg.mozilla.org/integration/mozilla-inbound/rev/71509cc11b9a

Comment 12

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/c80617d27cd9
https://hg.mozilla.org/mozilla-central/rev/18233d9d3ed5
https://hg.mozilla.org/mozilla-central/rev/71509cc11b9a

Comment 13

[Philip Chee]

comm-central bustage fix:
https://hg.mozilla.org/comm-central/rev/baddbdd06d16

Comment 14

[Boris Zbarsky [:bzbarsky]]

> comm-central bustage fix:

Oy, sorry about that.  :(