💾 Archived View for rawtext.club › ~sloum › geminilist › 007170.gmi captured on 2023-11-14 at 08:48:09. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2021-11-30)
-=-=-=-=-=-=-
Omar Polo op at omarpolo.com
Tue Sep 21 18:01:39 BST 2021
- - - - - - - - - - - - - - - - - - -
stack at tilde.club writes:
[...]
I am using BearSSL (for its minimalism). I started with the sample
provided, modified for the Gemini request. I keep getting stuck with
error 62 (probably because I don't understand how to deal with TLS).
I attempted to generate the CA boilerplate code, but there are some
ominous hints that the code is HTTPS-specific.
It's just a wild guess, I never touched BearSSL, but smell like apossible failure in verifying the certificates. In geminispace theusage of TOFU and self-signed certificates is way more popular than inthe web, and TLS libraries by defaults try to verify the certificateagainst the installed CAs. For the record, on libtls is the function istls_config_insecure_noverifycert.
regarding the other mail: the code in gemini.c for telescope <= 0.4 isprobably a bit easier to grasp. With telescope 0.5 gemini.c renamednet.c and handles gopher and finger too. Moreover, both telescope andgmid use async I/O and relies a lot on libevent, so the code may seem abit harder to follow at first given how it's splitted in more functions.
The code in the test program gg.c uses blocking I/O and is a singlefunction in contrast.
HTH
Omar Polo