275219 - nsIProfileChangeStatus.{idl,h} is not frozen but in the SDK

Status: RESOLVED INCOMPLETE

(Core Graveyard :: Profile: BackEnd, defect)

Description

[Christian :Biesinger (don't email me, ping me on IRC)]

attachment 120910 [details] [diff] [review] added this interface to the SDK, but it is not frozen. that
was checked in near 2003-05-02 for 1.4beta(-ish).

Comment 1

[Darin Fisher]

The profile change notifications are somewhat frozen, and there are applications
that depend on them.  As for nsIProfileChangeStatus, I'm not sure if anyone is
using that interface.  I'm tempted to try to freeze this observer API since it
doesn't seem like it would box us in too much.  Thoughts?  (Does the toolkit
profile stuff support all of this?)

Comment 2

[Benjamin Smedberg]

The toolkit profile code does not support nsIProfileChangeStatus, and I don't
intend to. You can do the same thing with the startup observer category and then
register a profile-change listener.

Comment 3

[Darin Fisher]

bsmedberg: are you sure?  what if that breaks embedders of gecko?  how can
libxul be an implementation of the GRE without implementing the GRE interfaces?
 i don't know for sure if anyone is using this, but we should treat this carefully.

Comment 4

[Benjamin Smedberg]

If the interface had been marked frozen, I probably would have implemented it in
the toolkit. But it's not. It was added for the mailnews offline-startup
extension, which I converted to use an app-startup category observer.

Comment 5

[Christian :Biesinger (don't email me, ping me on IRC)]

does toolkit at least support the observer topics from this file? 

(http://lxr.mozilla.org/seamonkey/source/profile/public/nsIProfileChangeStatus.idl)

Comment 6

[Benjamin Smedberg]

I support the ones that make sense since we don't do profile-switching:

at startup:
profile-do-change
profile-after-change

at shutdown:
profile-before-change

Comment 7

[Darin Fisher]

except that the veto API is not supported.  hmm... i guess that doesn't really
matter.  what is there really to veto.  anyways, we should freeze that subset
and include an appropriate header file in the SDK.

Comment 8

[Benjamin Smedberg]

Code gone. woot.