STINNER Victor <[EMAIL PROTECTED]> added the comment:
Here is a trace of Valgrind:
>>> b=bytearray(2)
>>> b.count("xxxx", 3493403, 0)
0
>>> b.count("xxxx", 23131230123012010231023, 0)
==13650== Invalid read of size 1
==13650== at 0x812718A: fastsearch (fastsearch.h:67)
==13650== by 0x81272CB: stringlib_count (count.h:22)
==13650== by 0x8128611: bytes_count (bytearrayobject.c:1198)
==13650== by 0x81376B6: PyCFunction_Call (methodobject.c:81)
==13650== by 0x80D913D: call_function (ceval.c:3679)
==13650== by 0x80D603A: PyEval_EvalFrameEx (ceval.c:2370)
==13650== by 0x80D7891: PyEval_EvalCodeEx (ceval.c:2942)
==13650== by 0x80D0A8F: PyEval_EvalCode (ceval.c:515)
==13650== by 0x80FB4FE: run_mod (pythonrun.c:1330)
==13650== by 0x80FA166: PyRun_InteractiveOneFlags (pythonrun.c:836)
==13650== by 0x80F9EEA: PyRun_InteractiveLoopFlags (pythonrun.c:756)
==13650== by 0x80F9DAC: PyRun_AnyFileExFlags (pythonrun.c:725)
==13650== Address 0x8444553a is not stack'd, malloc'd or (recently) free'd
==13650==
==13650== Process terminating with default action of signal 11 (SIGSEGV)
Trace of gdb:
0x0812718a in fastsearch (s=0x89b9fcaf <Address 0x89b9fcaf out of
bounds>, n=-2147483647,
p=0xb7b54274 "xxxx", m=4, mode=0) at Objects/stringlib/fastsearch.h:67
67 if (s[i+m-1] == p[m-1]) {
(gdb) where
#0 0x0812718a in fastsearch (s=0x89b9fcaf <Address 0x89b9fcaf out of
bounds>, n=-2147483647,
p=0xb7b54274 "xxxx", m=4, mode=0) at Objects/stringlib/fastsearch.h:67
#1 0x081272cc in stringlib_count (str=0x89b9fcaf <Address 0x89b9fcaf
out of bounds>,
str_len=-2147483647, sub=0xb7b54274 "xxxx", sub_len=4) at
Objects/stringlib/count.h:22
#2 0x08128612 in bytes_count (self=0xb7d025f0, args=0xb7b5075c)
at Objects/bytearrayobject.c:1198
(gdb) print w
$1 = 2147483645
(gdb) print m
$2 = 4
Oh, look at stringlib_count (..., str_len=-2147483647, ..., sub_len=4):
str_len is negative.
_______________________________________
Python tracker <[EMAIL PROTECTED]>
<http://bugs.python.org/issue3967>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
