Le mardi 07 juin 2005 à 15:59 -0700, Yen-Ju Chen a écrit :

> On 6/7/05, Quentin <[EMAIL PROTECTED]> wrote:
> 
> > I have fixed previosusly a segmentation fault in LCSegmentMerger.m with
> > the attached patch.
> 
>   Thanx. I applied it with some modification,
>   though I am surprised that it causes a segmentation fault.

What I have thought is [readers maxDocument] is returning a strictly
positive value, but [LCIndexReader setNorms:bytes:offset:] is currently
not  implemented then bytes is still empty when we call this method,
then the loop iteration on bytes is done on garbage memory.

> >  Few tests are failing too, like this one : LCPriorityQueue.m:206:
> > warning: msgUKTrue.fail
> 
>   There are chance that one of the LCPriorityQueue test will failed
>   because of the random number.
>   If you run it again, or later, it should pass.
>   It depends on random number.

ok.

I can know report another error : 
TestStore.m:139 msgUKIntsEqual.pass
ukrun: Uncaught exception NSFileHandleOperationException, reason:
attempt to close closed file

Quentin.

--
Quentin Mathé
[EMAIL PROTECTED]


Reply via email to