Bug 119142 – REGRESION(r151091) ASSERTION FAILED: m_readableData || m_writableData

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit Qt

[Allan Sandfeld Jensen]

http/tests/misc/drag-over-iframe-invalid-source.html asserts with 'ASSERTION FAILED: m_readableData || m_writableData'. The problem seems to be that the r151091 refactoring left us with an incorrect assert.

[Allan Sandfeld Jensen]

Created attachment 207523 [details]
Patch

[Jocelyn Turcotte]

Arun, what do you intend with this assert?

[Arunprasad Rajkumar]

(In reply to comment #2)
> Arun, what do you intend with this assert?

Either m_readableData or m_writableData should be valid, and not both.

I just moved the legacy to new interface, old code also had this assertion.

But anyhow I run all LayoutTests to confirm, Don't know how I missed this :(

[Arunprasad Rajkumar]

(In reply to comment #3)
> But anyhow I run all LayoutTests to confirm, Don't know how I missed this :(

I remember running all test case before raising for a review, whether that test case is added newly?(atleast after a commit build-bot might have run LayoutTests right?, i'm wondering how it is missed)

[Jocelyn Turcotte]

The assertion in this case is that "at least one of the two" is non-null when readData() is called. So if you just ported it I think we can just remove it then, I was wondering if you were intending that readData shouldn't be called in this case.
Thanks.

[Arunprasad Rajkumar]

(In reply to comment #5)
> The assertion in this case is that "at least one of the two" is non-null when readData() is called. So if you just ported it I think we can just remove it then, I was wondering if you were intending that readData shouldn't be called in this case.
> Thanks.

'ASSERTION FAILED: m_readableData || m_writableData', in this case both were null right? Is it expected?

[Allan Sandfeld Jensen]

(In reply to comment #6)
> (In reply to comment #5)
> > The assertion in this case is that "at least one of the two" is non-null when readData() is called. So if you just ported it I think we can just remove it then, I was wondering if you were intending that readData shouldn't be called in this case.
> > Thanks.
> 
> 'ASSERTION FAILED: m_readableData || m_writableData', in this case both were null right? Is it expected?

That is expected. The old assert you ported was a bit longer and tested another value you removed.

[WebKit Commit Bot]

Comment on attachment 207523 [details]
Patch

Clearing flags on attachment: 207523

Committed r153461: <http://trac.webkit.org/changeset/153461>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.