Freeze the following services/interfaces: nsIPrefService nsIPrefBranch Please follow the guidelines outlined in the "How to freeze an interface" at http://www.mozilla.org/projects/embedding/EmbedInterfaceFreeze.html to freeze these interfaces.

->0.9.6 Hi Brian : If you've done changing this interface, could we get this frozen... Thanks Chak

Just finished it up a few minutes ago. Doing a build right now to make sure nothing broke. :)

changing qa contact to Dharma.

I have attached a patch which adds documentation to the 2 interfaces we are freezing, as well as the remainder of the preferences interfaces which just needed this done anyway. There are also a couple of minor "correctness" changes that stemmed from the documentation effort. Input, reviews please.

Comment on attachment 52253 [details] [diff] [review] Changes for API freeze. sr=alecf

Comment on attachment 52253 [details] [diff] [review] Changes for API freeze. I think alec meant to check has-super-review. I checked it for him. has-review = valeski Looks great!

Comment on attachment 52257 [details] [diff] [review] New patch file that includes nsIPrefBranch.idl, and not 2 copies of nsIPrefBranchInternal.idl Sorry about my previous r=, I missed some stuff :-(. 1. Please remove the PREF_INVALID, LOCKED, REMOTE, LI_LOCAL, etc, and the associated comment. Those should live in the impl, not at the iface level. 2. Please use the attribute keyword in idl for the "get" "set" methods, when you can. This should cut down on description as well. boolean getBoolPref(...) becomes readonly attribute boolean boolPref(...) for example. I don't think there's a "writeonly" though, so those you'd have to keep as "setFoo(...)"

Comment on attachment 52257 [details] [diff] [review] New patch file that includes nsIPrefBranch.idl, and not 2 copies of nsIPrefBranchInternal.idl New patch posted based on valeski's comments (the ones that compiled anyway :D).

looks ok, sr=alecf On the constant values like PREF_INVALID and PREF_LOCKED, am I correct in assuming that the values are used by getPrefType()?

getPrefType() returns PREF_STRING, PREF_INT, or PREF_BOOL. It could potentially return 0 (PREF_INVALID) if the preftype is none of the above, but I'm not really sure how this could happen. I could leave PREF_INVALID in just in case... The PREF_LOCKED flag is basically returned via prefIsLocked(), the PREF_USER_SET value returned via prefHasUserValue(). PREF_CONFIG is set internally, but never used and PREF_REMOTE & PREF_LI_LOCAL are completely unused.

Checked in.

verified against Mozilla 1.1a Gecko 20020617 build. Verified patch checkin for nsIPrefBranch.idl, nsIPrefBranchInternal.idl, nsIPrefService.idl, nsISecurityPref.idl, nsPref.cpp, nsPrefBranch.cpp (only change is removal of NS_LITERAL_STRING in observer->Observe()), and nsPrefService.cpp