| commit | 5c85fc8552e8ce811833351b68b3d70b9bb6bf4f | |
| author | Miklos Vajna <vmiklos@collabora.com> | |
| Wed, 23 Nov 2022 07:13:44 +0000 (23 08:13 +0100) | ||
| committer | Miklos Vajna <vmiklos@collabora.com> | |
| Wed, 23 Nov 2022 08:06:33 +0000 (23 09:06 +0100) | ||
| tree | db8c369d2734eabc7035df1aca8ef05837bd0cf8 | treesnapshot (tar.gz zip) |
| parent | 7255d21f812b7333c83b4548e2b5b102adb606a9 | commitdiff |
exteral/freetype: improve -fsanitize=non-null-attribute fix
Quoting <https://gerrit.libreoffice.org/c/core/+/143098/2#message-62ac4499370dd88eeb3181f52d817e831150c94d>:
> What we do in other such cases is not make it conditional on the
> pointer argument being non-null, but rather make it conditional on the
> length argument not being zero. That way, it only catches the
> "harmless" case of an "irrelevant" null pointer in combination with a
> zero length, and doesn't hide bugs where the pointer is erroneously null
> while the length is non-zero.
So check for the length and not for the pointer argument.
Change-Id: I1894eb67d49a9ae40cdce29de865172400a271dd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/143134
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Quoting <https://gerrit.libreoffice.org/c/core/+/143098/2#message-62ac4499370dd88eeb3181f52d817e831150c94d>:
> What we do in other such cases is not make it conditional on the
> pointer argument being non-null, but rather make it conditional on the
> length argument not being zero. That way, it only catches the
> "harmless" case of an "irrelevant" null pointer in combination with a
> zero length, and doesn't hide bugs where the pointer is erroneously null
> while the length is non-zero.
So check for the length and not for the pointer argument.
Change-Id: I1894eb67d49a9ae40cdce29de865172400a271dd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/143134
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
| external/freetype/ubsan.patch | diffblobblamehistory |