Karlson2k <k...@narod.ru> writes: > Hi! > > First patch is direct fix for crash on Win64 (as result of invalid pointer > casting), also fix possible crash on any platform with > sizeof(size_t) != sizeof(long). > Second patch is similar fix for function that are not (yet) crashing. > Third patch is some kind of optimization and protection for buffer overruns - > once size of string is computed in any function, it will be reused in > subsequent called internal functions.
Hi! I am slowly returning to these emails. I think I was finally able to reproduce the problem you were seeing, and it was fixed in libidn version 1.42: https://lists.gnu.org/archive/html/help-libidn/2022-02/msg00000.html https://lists.gnu.org/archive/html/help-libidn/2024-01/msg00000.html I see that some of your patches were used but removed when upgrading to version 1.42: https://github.com/msys2/MINGW-packages/commit/9ddd87aeb6df905cf825c3624e9bd614515bc5f1 Are there any remaining concerns using libidn on Windows? We now build it in various configurations in GitLab CI/CD for Windows, so I'm hoping all of these problems are gone. /Simon
signature.asc
Description: PGP signature
