]> Gitweb @ Texas Instruments - Open Source Git Repositories - git.TI.com/gitweb - opencl/llvm.git/commit
[ADT][SmallVector] Flip an assert comparison to avoid overflows yielding false-negati...
authorAhmed Bougacha <ahmed.bougacha@gmail.com>
Wed, 7 Jan 2015 02:42:01 +0000 (02:42 +0000)
committerAhmed Bougacha <ahmed.bougacha@gmail.com>
Wed, 7 Jan 2015 02:42:01 +0000 (02:42 +0000)
commit8a19eb2a85d42f6976b57a7b5c1ec6fb6b8a88c7
tree69ed5592001c0b25df1fc4d2dff7d0f83add94df
parent8f1414458d0d7ce3238cbcf793b5e429013dfac3
[ADT][SmallVector] Flip an assert comparison to avoid overflows yielding false-negatives. NFC.

r221973 changed SmallVector::operator[] to use size_t instead of unsigned.

Before that, on 64bit platforms, when a large index (say -1) was passed,
truncating it to unsigned avoided an overflow when computing 'begin() + idx',
and failed the range checking assertion, as expected.
With r221973, idx isn't truncated, so the addition wraps to
'(char*)begin() - 1', and doesn't fire anymore when it should have done so.

This commit changes the comparison to instead compute 'end() - begin()'
(i.e., 'size()'), which avoids potentially overflowing additions, and
correctly triggers the assertion when values such as -1 are passed.
Note that the problem already existed before that revision, on platforms
where sizeof(size_t) == sizeof(unsigned).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225338 91177308-0d34-0410-b5e6-96231b3b80d8
include/llvm/ADT/SmallVector.h