On 20/05/2025 at 21:54, Holger Wansing wrote:
If I understand correctly, there are two ways, why those firmware files are copied/installed to the target system: - via some mechanism, when speech synthesis is used during installation (only loose firmware files are copied, no deb installed officially the apt way),
Not only when speech synthesis is used but whenever gtk/initrd.gz is used (by either graphical install or install with speech synthesis).
- via firmware-detect mechanism, when involved hardware is found (officially installing via apt). The second case should be solid enough to rely on, so why copy the files to /target in the first case at all?
See bug#992699 [1] for details. TL;DR:Speech synthesis needs sound to be enabled very early in the installation process, before the installation media which contains firmware packages is mounted.
[1] <https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=992699>