On Tue, Sep 26, 2023 at 08:12:43AM -0500, Andrea Bolognani wrote:
On Tue, Sep 26, 2023 at 01:14:33PM +0100, Daniel P. Berrangé wrote:
On Tue, Sep 26, 2023 at 07:02:19AM -0500, Andrea Bolognani wrote:
I don't think it helps much with not storing additional data inside the build system, unless we want to store the contents of the various common snippets in separate files? Something like
common_service = fs.read('common_service.inc') unit_conf = configuration_data({ 'common_service' = common_service, })
We'd have to fake fs.read() because it was introduced in 0.57 though. And we'd have to run the contents of the common parts through variable substitution anyway, because they will contain a bunch of lines like
Also=@service@.socket Also=@service@-ro.socket Also=@service@-admin.socket
I'm not sure the result would look much better, but I can give it a try.
Don't try to do any of this in meson. We should just have a standalone python script that can combine the daemon specific unit file contents with the common unit file contents. eg
scripts/merge-unit-file.py \ src/qemu/virtqemud.service.in \ src/rpc/virtd.service.in \ build/src/virtqemud.service
It feels a bit silly to shell out to Python to perform what is ultimately a bunch of variable substitutions, as if that wasn't part of Meson's core feature set... But I'll give it a try and see how it turns out.
IMHO Meson's job is to control the build process, rather than to actually be the build process. I think of this as "compiling" the unit files and the python sript is our compiler, which meson is to control.
Can you please take a look at the remaining patches in the meantime, and provide feedback on the changes that are made to the various services and sockets as part of them? Thanks in advance :)
With regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|