#35 Correct maximum size for a couple of spins
Merged by bcotton. Opened by adamwill.
fedora-pgm/ adamwill/pgm_docs 202212-sizes-update  into  main

Download 35.patch

In both of these cases, the maintainer actually decided on a larger maximum in the F31 cycle and updated the wiki page (which was used at that time), but the change was lost when the wiki page for the next release was made (presumably the program manager based it on the original version of the F31 page). To respect the maintainers' wishes we should correct these sizes. relval has been checking for these sizes all along, not the ones currently listed on the page.

Pull-Request has been merged by bcotton

Perhaps I should just drop the sizes from these pages. Since relval takes actual action based on the sizes it knows about, it makes sense to keep that and drop this.

well, the way it's "supposed to work" is that the pages are the canonical definition and relval reflects those. I don't let people get the sizes changed in relval's definition without them being updated on the pages first. But updates to the pages need to be carried forward :D

relval's definitions live here, but they aren't much use as a canonical public record, they're very tied to how relval works (for a start, there's a catch-all default of 4.7G which is in the code, not that file). I wouldn't want to ditch the human-readable definitions if that's the only other record we have.

also, ugh, I just realized these should have been GB, not GiB. I'll send a follow-up.

Follow-up: https://pagure.io/fedora-pgm/pgm_docs/pull-request/36

well, the way it's "supposed to work" is that the pages are the canonical definition and relval reflects those. I don't let people get the sizes changed in relval's definition without them being updated on the pages first. But updates to the pages need to be carried forward :D

This is a topic for broader, later discussion, but I wonder if there's much value in having a canonical definition. I will argue with myself about that for a while. Thanks for the fixes.

I would say we want at least the release-blocking image size limits, as they really are key 'distro policy' - that is where we say it really matters to Fedora that the netinst images are CD size and the DVD images are...DVD size. If we don't have those pages we don't have a logical backing for blocking releases on images being oversize.

The ones for non-blocking images are much more in the line of conveniences for the maintainers, really. They, at least some of them, want to know if their images suddenly blow up way bigger than they're intended to be. Those ones we could maybe stop documenting and just handle between relval and bugzilla and the maintainers, with maybe just a generic note about how size checking is handled on the Spins page.

Metadata