Re: [PATCH qemu 1/2] generate and ship static list of CPU models

[Fiona Ebner <f.ebner@proxmox.com>]

Am 11.02.26 um 11:32 AM schrieb Thomas Lamprecht:
> Am 10.02.26 um 14:53 schrieb Fiona Ebner:
>> Fail the build when the list changes, so that new models can be
>> picked up selectively and other changes that might require adaptations
>> are not missed.
> 
> thanks, any objection in storing this as pretty-formatted json?
> As then the default line-based git diff is making comparisons simpler,
> as while word-diff is great, it has it's limits IMO. 

No objection, I was considering it too after sending and seeing the very
long line in the patch. But wasn't sure :)

> No need for a v2 in any case, as that change is easy enough to squash in.

Thanks!

> btw. starting to wonder if we should produce an extra architecture "all"
> packages for such files, could reduce churn on selective uploads; but I'd
> figure that the current rate of change will slow down a bit soon.

One file for recognized aarch64 CPU flags might still come along in the
context of the custom CPU model edit API/UI. Otherwise, changes to those
files usually come together with switching to a new upstream version.