Closed
Bug 108907
Opened 24 years ago
Closed 17 years ago
Freeze XMLHttpRequest interface
Categories
(Core :: XML, defect, P2)
Core
XML
Tracking
()
RESOLVED
WONTFIX
Future
People
(Reporter: hjtoi-bugzilla, Assigned: hjtoi-bugzilla)
Details
Attachments
(1 file)
|
4.08 KB,
patch
|
Details | Diff | Splinter Review |
Must freeze this interface to let developers rely on it. Since it is copied from
a stable Microsoft interface should be relatively easy. Must get rid of
nsIChannel from the interface, and freeze all other interfaces that are required
by this interface as well as the ones those require and so forth. All in all
maybe about 10 DOM interfaces plus this one.
| Assignee | ||
Updated•24 years ago
|
Priority: -- → P2
Target Milestone: --- → mozilla0.9.8
| Assignee | ||
Updated•24 years ago
|
Target Milestone: mozilla0.9.8 → mozilla0.9.9
| Assignee | ||
Comment 1•24 years ago
|
||
Need to freeze:
nsIXMLHTTPRequest
nsIOnReadystatechangeHandler
nsIJSXMLHttpRequest
http://lxr.mozilla.org/mozilla/source/extensions/xmlextras/base/public/nsIXMLHttpRequest.idl
nsIDOMDocument [status FROZEN!]
http://lxr.mozilla.org/mozilla/source/dom/public/idl/core/nsIDOMDocument.idl
nsIDOMEventListener [status FROZEN]
http://lxr.mozilla.org/mozilla/source/dom/public/idl/events/nsIDOMEventListener.idl
nsIVariant
nsIDataType
(nsIIDPtr, nsQIResult, voidPtr, nsIID, nsID ?)
http://lxr.mozilla.org/mozilla/source/xpcom/ds/nsIVariant.idl
Need to get rid of nsIChannel. I think the only user was the IBM P3P code. Am I
right?
MS has two more members, responseBody and responseStream. Not sure what to do
with them yet.
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/xmlsdk30/htm/xmobjxmlhttprequest.asp
| Assignee | ||
Comment 2•24 years ago
|
||
| Assignee | ||
Comment 3•24 years ago
|
||
valeski gave me this URL as a poor man's guide to what is required for an
interface to become FROZEN:
http://www.mozilla.org/projects/embedding/EmbedInterfaceFreeze.html
He also noticed that the name should change to nsIXMLHTTPRequest (notice all
upper case HTTP). I am not too happy about that, but...
nsIVariant is in need of some documentation.
Interfaces should go one in each file.
| Assignee | ||
Comment 4•24 years ago
|
||
Adding valeski & vidur... I have a couple of questions to valeski and vidur
might have something to add as well.
1. Is it an absolute must that all interfaces be in separate files? In the case
of XMLHttpRequest there are three, but from JS there is logically just one (a
combination fo them all). nsIVariant also has several that in my opinion belong
into one file, except maybe for nsIWritableVariant.
2. I see changing the name very problematic at this point: it shipped with
Netscape 6.1 so there are likely to be sites using this. Should we make it a
warning in strict JS to use the old style, and also introduce XMLHTTPRequest? I
am not even sure if this is possible...
Finally a note to myself, all Mozilla extensions to XMLHttpRequest should be
clearly marked as such with comments.
Status: NEW → ASSIGNED
Comment 5•24 years ago
|
||
1. one iface per file isn't a hard and fast rule, rather, its a good guidline to
follow. if there are 'n' ifaces in a single .idl file, they at least needs to be
logically related, and the chances of someone needing to use all of them
together should be pretty high. the point is, if I only need nsIFoo, I don't
even want to have to see nsIBar (unless I really need to).
2. If the name change is too painful, let's avoid it.
| Assignee | ||
Comment 6•24 years ago
|
||
It seems there is pressure to still change nsIVariant. It is also relatively new
interface, so it seems to be unclear when it would be ready to be frozen.
Eventually it should be, as I think it is one of our "fundamental data types",
so to speak.
This got me thinking that we currently do not have a process to freeze the
interfaces that JS only sees (and not worry about C++). I don't know if there
should be. It would be a lot easier to freeze XMLHttpRequest's "JS interface".
I am starting to think this freeze should happen post-1.0...
| Assignee | ||
Comment 7•24 years ago
|
||
Later, although I might remove nsIChannel anyway...
Updated•23 years ago
|
QA Contact: petersen → rakeshmishra
Updated•22 years ago
|
QA Contact: rakeshmishra → ashishbhatt
Comment 8•22 years ago
|
||
Please consider bug 218236 when deciding whether to remove nsIChannel. The only
workaround I could find is checking the channel every second - there is no other
way to know that the request is completed.
Lets not do this. We're planning on no longer freezing interfaces, and this one is still having features added to it.
Status: ASSIGNED → RESOLVED
Closed: 17 years ago
Resolution: --- → WONTFIX
You need to log in
before you can comment on or make changes to this bug.
Description
•