irwir

> In TF-PSA-Crypto, we intend to only support CMake as the build system. In particular, `CMakeLists.txt` is the only build script. Presumably Mbed TLS will follow suit It appears that...

Is UPnP enabled in the device?

Just double checking. It might be worth testing the router with PCP (NAT/PMP) implementation, such as [this library](https://github.com/libpcpnatpmp/libpcpnatpmp).

@miniupnp > the XML root description of your UPnP device is incorrect. No UPnP client can properly work with it as the informations about services are missing. Unlike `miniupnpc`, the...

I do not know this router (could it be a clone of `Huawei H151-370` model?), and thus making a wild guess. The device reply allows to guess that UPnP is...

According to the [crypto++](https://cryptopp.com/) site, the current version 8.9.0 could be built with Visual Studio 2003 - 2022. The issue #1250 makes references to much older #123 and #496. Building...

This PR essentially does not fix the issue, because zero count in `memcpy` never was a trouble. The problem is `ELEM_MAX` being larger than clang allows.

> * Use seed.data() instead of pointer arithmetic for static analyzer `seed.data()` is a pointer, hence `seed.data()+4` is still pointer arithmetics. > * Add CRYPTOPP_ASSERT to verify bounds explicitly There...

> The benefit is that using an explicit `.data()` pointer helps Clang/Fortify track buffer bounds in static analysis, which reduces false positives in practice. Exactly the opposite is true. Abstract...

This complicates code but does not resolve the isssue: 32-bit signed integer vs unsigned 32 and 64-bit requires type casts with potential loss of data. Proper fix should restore the...