guix-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[bug#63877] [PATCH v2] gnu: services: web: Allow specifying extra php-fp


From: Timo Wilken
Subject: [bug#63877] [PATCH v2] gnu: services: web: Allow specifying extra php-fpm environment variables.
Date: Sun, 15 Oct 2023 22:54:06 +0200

Hi Bruno, (hi Ludo'), thank you for your detailed feedback and sorry for not
responding earlier!

On Mon Jun 5, 2023 at 5:44 AM CEST, Bruno Victal wrote:
> Ungexp-ing lists can be rather tricky [...]
>
> You need to quote the list [...]

I was thinking of something closer to the example I added to doc/guix.texi in
my patch. The gexp would not be a list directly, but instead be some code that
would produce a list when evaluated, e.g.:

--8<---------------cut here---------------start------------->8---
#~(list (string-append "SSL_CERT_DIR=" #$nss-certs "/etc/ssl/certs"))))
--8<---------------cut here---------------end--------------->8---

That would let you refer to store paths in variable values, instead of being
limited to literal strings.

As far as I know, the following throws an error, and `file-append' instead of
`string-append' wouldn't work because of the `"SSL_CERT_DIR="' prefix, right?

--8<---------------cut here---------------start------------->8---
#~(#$(string-append "SSL_CERT_DIR=" nss-certs "/etc/ssl/certs"))))
--8<---------------cut here---------------end--------------->8---

If you have any ideas on a better way to do this, let me know!

> Bonus points if you can write a small system test for this. (see
> gnu/tests/web.scm for inspiration)
> For our purposes, a pair of HTTP servers where one of them uses a
> self-signed certificate will suffice.

Thanks for the pointer! I'll try to get something basic working along the
lines of the php-fpm tests already there, and send a PATCH v3 soon. I was
thinking of only verifying that an arbitrary sentinel variable is set, and not
bother to test SSL_*-related behaviour, but I can try to get the latter
working if you think that would be better.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]