When you want a cryptography library for #guile, you might be pushed to use guile-#gcrypt. I advise you to use #gnutls instead: since guile use gnutls for its web stack, it might save you a dependency, while adding lots of cool features (e.g. parsing PEM keys, a requirement for ActivityPub interoperability with Mastodon).
For instance, here is how #goblins could use it:
https://gitlab.com/spritely/guile-goblins/-/merge_requests/181
#guile #gcrypt #GnuTLS #goblins
If you are afraid to review my gnutls-guile merge requests because there is too much change, you can start with this one:
#GnuTLS #guile #crypto #cryptography #guix
Dear gnutls guilers,
I still have a few merge requests waiting for your comments:
#guile #GnuTLS #scheme #crypto #cryptography #guix
Dear gnutls guilers,
I still have a few merge requests waiting for your comments:
#guile #GnuTLS #scheme #crypto #cryptography
Hello! This is a gentle reminder that I’m still looking for feedback on my GnuTLS - Guile merge requests :
#GnuTLS #guile #crypto #cryptography
J’ai vu passer ceci :
https://thephd.dev/to-save-c-we-must-save-abi-fixing-c-function-abi
C’est la première fois que j’entends ce son de cloche vis-à-vis de realloc, comme un échec de standardisation.
Mais pour avoir croisé la route de GnuTLS récemment, la solution indiquée me semble un peu tirée par les cheveux non ? Il suffit de garder les vieilles fonctions et de rajouter un numéro de la honte derrière, comme nom_de_fonction2 (). J’ai loupé un truc ?
Dear gnutls guilers, I have a few more merge requests for you to examine!
#guile #GnuTLS #crypto #cryptography
That’s it, I have an initial merge request for HMAC in the #guile binding of #gnutls.
https://gitlab.com/gnutls/guile/-/merge_requests/5
Can someone there run the pipeline?
I’m not sure I can go further with my #gnutls bindings in pure scheme for #guile. The fact that gnutls expects me to pack arrays of structs, pass enumeration types as output parameters, and that enumerations may be extended in the future, makes me reconsider whether I can really do everything in scheme.
Finally, the most reasonable thing to do would be to further develop the existing C bindings.
So, I managed to get the start of a #guile binding to #gnutls. Compare it to the official gnutls binding:
- the official one does not have the functions I want;
- the official one is written in C, not in Scheme;
- the official one uses smobs and (as far as I understand) GC finalizers, mine uses explicit memory management with dynamic-wind.
Now, the next task is to replace guile-gcrypt by incorporating the missing functions, and enjoy the wider API.
It's #GNUSpotlight time! Fifteen new GNU releases in the last month: https://u.fsf.org/3ue #bash #emacs #gnugamma #gnutls #gnusasl and more -- hurrah! Thanks again, @bandali0 @bandali and big thanks to all the devs and other contributors!
#GNUSpotlight #bash #emacs #gnugamma #GnuTLS #gnusasl
Well, I'm back at attempting to get a #Heroku buildpack working for #Guile #Artanis (well, technically, getting #GnuTLS working with the Guile and Artanis buildpacks I already got working so you can use HTTPS with Artanis).
And, while I'm here, might as well update my Guile buildpack to 3.0; looks like it's working, so far…
#Heroku #guile #artanis #GnuTLS
If anyone wants to take a crack and help me out, I'm trying to build #GnuTLS with #Guile bindings in a #Heroku buildpack.
The project uses this buildpack I made to install Guile: https://github.com/WammKD/heroku-buildpack-guile-scheme.
That works fine and then I have this other one to try and build GnuTLS: https://github.com/WammKD/heroku-buildpack-gnutls-with-guile-bindings.
Unfortunately, that one's throwing an error of "g++: fatal error: -fuse-linker-plugin, but liblto_plugin.so not found" and I know too little to figure out why.