[Python-checkins] python/dist/src/Objects listobject.c,2.108,2.109
sliceobject.c,2.12,2.13 stringobject.c,2.165,2.166 tupleobject.c,2.64,2.65
unicodeobject.c,2.151,2.152
Michael Hudson
mwh@python.net
Tue, 11 Jun 2002 14:33:42 +0100 (BST)
- Previous message: [Python-checkins] python/dist/src/Objects listobject.c,2.108,2.109
sliceobject.c,2.12,2.13 stringobject.c,2.165,2.166
tupleobject.c,2.64,2.65 unicodeobject.c,2.151,2.152
- Next message: [Python-checkins] python/dist/src/Doc/api concrete.tex,1.15,1.16
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]
On Tue, 11 Jun 2002, Neal Norwitz wrote:
> Michael:
>
> There is a problem with the patch below.
> I think it is a general problem (for all types).
> I verified w/lists only. I'll describe the problem inline below.
>
> >>> n = [1, 2, 3]
> >>> n[10000:30000:2]
> Segmentation fault (core dumped)
Ouch!
[...]
> Here is where the problem is. You are referencing self list
> with cur, but cur is not guaranteed to be < list length.
> I think all you have to do is add another condition to the for loop
> && cur < PyList_Length(self) (or should that be PyList_Size(self)?).
No, PySlice_GetIndicesEx was wrong. It was returning a slicelength of 1
when it should have been 0. Fix imminent.
Cheers,
M.
- Previous message: [Python-checkins] python/dist/src/Objects listobject.c,2.108,2.109
sliceobject.c,2.12,2.13 stringobject.c,2.165,2.166
tupleobject.c,2.64,2.65 unicodeobject.c,2.151,2.152
- Next message: [Python-checkins] python/dist/src/Doc/api concrete.tex,1.15,1.16
- Messages sorted by:
[ date ]
[ thread ]
[ subject ]
[ author ]