368328 - ASSERTION: Must have a buffered input stream: 'NS_InputStreamIsBuffered(pIStream)'

Status: RESOLVED FIXED

(Core :: XPCOM, defect)

Description

[Jeff Walden [:Waldo]]

See URL in a debug build.

The assertion seems to occur because NS_InputStreamIsBuffered tries to read a byte from a stream which is empty and fails in some particular way -- a non-empty string won't trigger the assertion.

Comment 1

[Boris Zbarsky [:bzbarsky]]

I'd say this is a bug in NS_InputStreamIsBuffered(), no?  It's lying about the stream...

Comment 2

[Jeff Walden [:Waldo]]

Yeah, probably; I was a bit lazy in researching the bug when I filed it.  :-\

Comment 3

[Ryan Jones-Ward [:sciguyryan]]

Attachment: Patch v1

Patch v1

I'm not actually sure we still need to |result| here but this should check for empty streams.

Comment 4

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

Comment on attachment 264665 [details] [diff] [review]
Patch v1

sr=biesi, although it is unfortunate that this will not detect certain kinds of unsupported ReadSegments

Comment 5

[Ryan Jones-Ward [:sciguyryan]]

(In reply to comment #4)
> (From update of attachment 264665 [details] [diff] [review])
> sr=biesi, although it is unfortunate that this will not detect certain kinds of
> unsupported ReadSegments
> 

Maybe a follow-up bug could be filed to add a new function with the old checking abilities?

Comment 6

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

who would call it?

Comment 7

[Ryan Jones-Ward [:sciguyryan]]

I'm not sure but the option would still be there to use it, then again if it has no callers would it be worth it?

Comment 8

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

well really how would a caller choose between the two?

Comment 9

[Ryan Jones-Ward [:sciguyryan]]

(In reply to comment #8)
> well really how would a caller choose between the two?
> 

They couldn't really. The both would have basically the same functionality in most circumstances and only differ in a small number of situations...

Comment 10

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

right, so I don't think there's much point in adding that function.

Comment 11

[Boris Zbarsky [:bzbarsky]]

Checked in.  Thanks for the patch!

Comment 12

[Jeff Walden [:Waldo]]

If this is done as an xpcshell test, the assertion will actually cause the test to fail, if/when we get around to running the unit tests on debug builds in addition to unpackaged release builds.