| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| 0001-TODO-tests-for-base64-decoding-failures.patch | [PATCH 1/3] TODO tests for base64 decoding failures When the input is a suitably long string (24 to 26 characters), sv_to_uuid() decodes it as base64 but doesn't check if the result makes sense. The decoding process silently ignores illegal base64 characters and padding after '='. This can break test 28 when Perl pointers stringify to a suitably long string, such as "Blah=HASH(0x555555f30d18)". Add TODO tests showing the behaviour on all platforms. |
Niko Tyni <ntyni@debian.org> | yes | debian upstream | 2016-11-27 | |
| 0002-Check-that-a-base64-decoded-string-is-long-enough-to.patch | [PATCH 2/3] Check that a base64 decoded string is long enough to be a UUID This fixes test failures on platforms where Perl pointers stringify to a suitably long string, making "Blah=HASH(0x555555f30d18)" a candidate for base64 decoding. |
Niko Tyni <ntyni@debian.org> | yes | debian upstream | 2016-11-27 | |
| 0003-Fix-use-of-uninitialized-memory-in-new_uuid-when-ver.patch | [PATCH 3/3] Fix use of uninitialized memory in new_uuid() when version==1 This was discovered by running the test under valgrind. |
Niko Tyni <ntyni@debian.org> | yes | debian upstream | 2016-12-06 |