102779 - data: URLs never fire onload handlers

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

data: URLs don't fire onload handlers.  I suspect this has something to do with
something at the network level not firing.  This makes it difficult to put data:
URLs into a performance test for layout performance on very simple pages.

Comment 1

[Charles Manske]

Any plans to fix this?

Comment 2

[neeti]

darin.

Comment 3

[Darin Fisher]

this wouldn't be solved in necko...

-> layout

Comment 4

[Johnny Stenback (:jst)]

Rick, you're gonna love this one :-)

Comment 5

[rpotts (gone)]

Attachment: Patch to cause OnLoad events to fire -- and other cleanup

Comment 6

[rpotts (gone)]

how do people feel about getting rid of the nsIDataChannel interface?  Currently
it is identical to nsIChannel  !!

it seems like unnecessary overkill...  if people agree, i'll put together
another patch to nuke it ;-)

-- rick

Comment 7

[Darin Fisher]

Comment on attachment 61920 [details] [diff] [review]
Patch to cause OnLoad events to fire -- and other cleanup

patch looks good to me...  looks like you got all the different cancelation
paths handled. r/sr=darin

Comment 8

[Darin Fisher]

right!  nsIDataChannel is useless, and certainly could/should go away.

Comment 9

[Johnny Stenback (:jst)]

Comment on attachment 61920 [details] [diff] [review]
Patch to cause OnLoad events to fire -- and other cleanup

> NS_IMETHODIMP
> nsDataChannel::GetName(PRUnichar* *result)
> {
>-    NS_NOTREACHED("nsDataChannel::GetName");
>-    return NS_ERROR_NOT_IMPLEMENTED;
>+    nsXPIDLCString name;
>+
>+    if (mUrl) {
>+        mUrl->GetSpec(getter_Copies(name));
>+    }
>+    *result = ToNewUnicode(name);

Do we need any UTF8 fu here?

[...]
>+    return mStatus;
>+}
>\ No newline at end of file

Add a newline at the end of nsDataChannel.cpp, some odd ports (HP-UX?) will go
red if you don't add this.

Other than that, r/sr=jst

Comment 10

[rpotts (gone)]

patch checked in...

Comment 11

[Christopher Hoess (gone)]

testcase now works.  resolving fixed.

Comment 12

[Christopher Hoess (gone)]

*** Bug 102911 has been marked as a duplicate of this bug. ***