search

quicklisp / quicklisp-client

Quicklisp client.
http://www.quicklisp.org/
MIT License
298 stars 74 forks source link

hook into ASDF so that missing systems are dl'd by ql #207

Open attila-lendvai opened 3 years ago

attila-lendvai commented 3 years ago

it would by nice if the image could be globally sideeffected once (e.g. by installing an ASDF hook), and then later on any missing system would invoke quicklisp to attempt to download it.

using a toplevel ql:quickload call is convenient in most use-cases, but sometimes we cannot modify/customize some code deep down somewhere to call quickload, especially if we want to keep quicklisp optional.

example use case: a build.sh script that works both on a completely fresh linux install, and also in the customized environment of a developer, where it's only downloading from ql that which is not already installed/modified in the developer's environment.

there's a lengthy discussion about the technical details of this at https://github.com/quicklisp/quicklisp-client/pull/122#issuecomment-160419822. see this comment/discussion for a sketch of such a *system-definition-search-functions* hook.

quicklisp commented 3 years ago

What ASDF introduces that hook?

On Thu, Apr 15, 2021 at 10:00 AM Attila Lendvai @.***> wrote:

it would by nice if the image could by sideeffected once, e.g. by installing an ASDF hook, and then later on any missing system would invoke quicklisp to attempt to download it.

ql:quickload is convenient in most use-cases, but sometimes we cannot modify/customize some code deep down somewhere to call into quickload, especially if we want to keep quicklisp optional.

example use case: a kind of build.sh that works both on a completely fresh linux install, and also in the customized environment of a developer, where it's only downloading from ql that is not already installed/modified in the developer's environment.

there's a lengthy discussion about the technical details of an asdf hook at #122 https://github.com/quicklisp/quicklisp-client/pull/122

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/quicklisp/quicklisp-client/issues/207, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACPNLPBTAMJGCKTBY6J423TI3WRNANCNFSM427SBOLA .

attila-lendvai commented 3 years ago

@quicklisp ehh, sorry. not only did i read that comment sloppily, that wasn't even Fare's comment, but mine, and a hypothetical one at that.

i'll update/reword the issue description now.