GET requests on github API and raw links return bad host (status code: 0) about cpr

[whoshuu]

Just tried the example in the readme documentation again, and it works just fine. Can you give me a minimal snippet of the request that's bugging out for you?

[nabijaczleweli]

@IngCr3at1on OS, curl version?

[IngCr3at1on]

@whoshuu will have to get that later. @nabijaczleweli this occurs in Arch Linux with curl 7.45

[vitorhnn]

I'm having the same issue, but I found out that it happens only if you don't use your system's libcurl and it's just https requests, http works fine. I'm also on Arch Linux with curl 7.45 and OpenSSL 1.0.2.d

[IngCr3at1on]

@vitorhnn aye I've noticed it only seems to happen on https also, I forgot to note that in the initial issue (since all github api and raw links are handled over https I didn't feel this was necessary). I have the same version of OpenSSL also.

As for a snippet of the request it's literally the same one as the example that I was playing with.

@vitorhnn can you elaborate on what you mean by "don't use your system's libcurl"; more specifically was there something you were able to do to make it work?

[whoshuu]

@vitorhnn, I think the issue you're seeing is related to #31, which has yet to be implemented when building the included curl library. It makes sense that your system curl works just fine, since it's correctly using OpenSSL from your system as well. When #31 is implemented, let's circle back on this issue and see if it's resolved.

[whoshuu]

@IngCr3at1on, there's an option to either use your system's libcurl or the included one. The option is USE_SYSTEM_CURL (found here) and by default it's OFF. You can turn it on by setting an environment variable (ie, export USE_SYSTEM_CURL=ON), or setting a CMake definition cmake -DUSE_SYSTEM_CURL=ON .. && make.

Additionally, the option should be exposed in the CMake gui if that's what you're using to configure your project.

[vitorhnn]

@IngCr3at1on, set USE_SYSTEM_CURL to ON on your CMakeLists.txt, like here

[IngCr3at1on]

Alright @whoshuu and @vitorhnn I'll try that out. Assuming it works I'll close this being that #31 already exists. Thanks.

[nabijaczleweli]

@IngCr3at1on And?

[COM8]

Closing this, since it does not seem to be relevant any more.