Matt Caswell

While testing this out I hit an assertion failure on this line in fuzz/pem.c: ```` OPENSSL_assert((size_t)BIO_write(in, buf + 1, len - 1) == len - 1); ```` My temporary solution...

I have been unable to reproduce any of these 3 new detections. However I believe I have a fix in #21741 for [this one](https://github.com/openssl/openssl/pull/21668#issuecomment-1675181186) and [this one](https://github.com/openssl/openssl/pull/21668#issuecomment-1676352498). For [this one](https://github.com/openssl/openssl/pull/21668#issuecomment-1676039342)...

Thanks. That was enough for me to track this down. Fix in #21753.

Fix for the latest issue in #21796

> but due to the ignored memory allocation error, we now have a unique setup with just one crypto algorithm missing. And the fuzzing vector has not yet been examined...

Fix for that in #22536

#22536 was merged...

It looks to me like this bit of code could simply be deleted: ```` C #if !defined(OPENSSL_NO_ENGINE) && !defined(FIPS_MODULE) if (src->engine != NULL && ENGINE_init(src->engine) == 0) goto err; ret->engine...

> This does not mention that potentially two such encodings can be given, which then need to be concatenated. I don't see any concatenation in the above code? It just...

> Some even hand-craft ASN.1 encodings We do this too...well the code to do it is auto-generated. See the various code files here: https://github.com/openssl/openssl/tree/master/providers/common/der The reason being is that the...