Re: new egg: http-curl

Mario Domenech Goulart via Chicken-users <[email protected]> Mon, 01 Jun 2026 19:18:36 +0200
Newsgroups gmane.lisp.scheme.chicken
Message-ID <[email protected]>
Hi Rolando,

On Sun, 31 May 2026 21:08:10 -0700 Rolando Abarca <[email protected]> =
wrote:

> I still don't see the egg in the registry, can you please double check? i=
f there's anything broken on my side, happy to
> fix it!

Sorry -- my bad.  By mistake I added the link to the HTML-formatted
.release-info file to the egg-locations file.

Peter has fixed that and the egg is now be available.

All the best.
Mario

> On Thu, Apr 30, 2026 at 6:03=E2=80=AFPM Rolando Abarca <cpm.rolandoa@gmai=
l.com> wrote:
>
>  I just bumped the version and gated the network tests behind an env var.=
 They don't run by default.
>
>  Thanks again,
>  Rolando
>
>  On Wed, Apr 29, 2026 at 12:55=E2=80=AFPM Mario Domenech Goulart <mario@p=
arenteses.org> wrote:
>
>  Hi Rolando,
>
>  On Tue, 28 Apr 2026 21:32:20 -0700 Rolando Abarca via Chicken-users <chi=
[email protected]> wrote:
>
>  > I've been testing a new egg that's a drop-in replacement for http-clie=
nt, http-curl:
>  > https://forgejo.rolando.cl/cpm/http-curl that, like the name implies, =
uses libcurl as the transport. The main
>  motivation
>  > was to have an http-client library that could have better https suppor=
t (I kept having lots of issues with
>  http-client in
>  > oauthtoothy because it failed to connect to the APIs when trying to ex=
change the token).
>  >
>  > The goal for this egg was to make it as easy as possible to transition=
 and maintain the same API as
>  http-client. You need
>  > libcurl to build the egg.
>  >
>  > repo: https://forgejo.rolando.cl/cpm/http-curl
>  > release-info: https://forgejo.rolando.cl/cpm/http-curl/raw/branch/main=
/http-curl.release-info
>  > docs: https://forgejo.rolando.cl/cpm/http-curl/src/branch/main/README.=
md
>  >
>  > I would appreciate it if it was added to the coup.
>
>  Thanks.  Your egg has been added to CHICKEN 5.
>
>  I'd have a request, though: could you please reconsider the decision of
>  using external resources for tests?  Some reasons:
>
>  * They might introduce sources of indeterminist to tests.
>
>  * Tests will not succeed in scenarios without network access.
>
>  * Tests might fetch malicious content if external resources get
>    compromised.
>
>  All the best.
>  Mario
>  --=20
>  https://parenteses.org/mario
>

--=20
https://parenteses.org/mario