From 3e54ecd73665934187932e61efe922a472887246 Mon Sep 17 00:00:00 2001
From: Pjotr Prins
Date: Thu, 26 Jun 2025 17:01:01 +0200
Subject: Prepare for ninja

---
 doc/code/pangemma.md | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

(limited to 'doc/code')

diff --git a/doc/code/pangemma.md b/doc/code/pangemma.md
index b588b7e..013816a 100644
--- a/doc/code/pangemma.md
+++ b/doc/code/pangemma.md
@@ -209,6 +209,8 @@ There are several methods for sharing data between threads and processes. First
 The build system for development relies on Guix. See the header of [guix.scm](../guix.scm).
 Guix handles depencies well, so it is a great base to work from.
 
-For the actual build system we'll probably roll our own.
+Build systems have to target different platforms. pangemma will mostly be run on Linux on AMD64, but there are always people who want more and it requires support from the build system. Also we may need targets for debug and architecture optimized builds. Some of that may be handled through Guix.
+
+For the actual build system we may roll our own.
 All build tools I know are too messy, including CMake, meson and make itself. So, as a real hacker, we
-reinvent the wheel. A starting point is Arun's [g-exp blog](https://systemreboot.net/post/g-expressions-makings-of-a-make-killer).
+can reinvent the wheel. One starting point is Arun's [g-exp blog](https://systemreboot.net/post/g-expressions-makings-of-a-make-killer) though it would require a non-Guix way of building things.
-- 
cgit v1.2.3