On Wed, 2019-04-17 at 17:58 +0100, Daniel P. Berrangé wrote: > On Wed, Apr 17, 2019 at 06:38:55PM +0200, Andrea Bolognani wrote: > > Actually, the above is not quite correct: even if an application > > wanted to build the display string from scratch, they would not be > > able to do so because there is no field anywhere containing the > > strings "MacOS X", "Debian" or "Ubuntu". > > That's by design, because the product name is not a simple > concatenation of a base name + version number. It is essentially > an opaque string in its own right, arbitrarily decided by the > vendor. You could argue that the vendor-chosen for the latest Debian release is "Debian 9 (stretch)" based on https://www.debian.org/releases/ and similarly for Ubuntu, but it's definitely not as clear-cut as it might be with the likes of Windows and macOS :) > > There's one more detail to consider: even if the application was > > willing to construct the display string itself, how would it know > > that it's "Ubuntu Server 18.04 LTS" but "Fedora 29 Workstation"? > > They'd have to write code for it, and end up not handling all cases > > correctly... If we just provide the information is osinfo-db, on > > the other hand, it's just a quick API call away. > > The application can just display <name> field as is, > either the top level name from the OS, or the variant > name if they have identified a specific varaint. > > They can optionally append a codename string to that > when displaying it. That would result both in "Debian 9 (stretch)" (good) and "macOS Mojave (Mojave)" (bad), though. So we should probably not recommend applications do that. -- Andrea Bolognani / Red Hat / Virtualization _______________________________________________ Libosinfo mailing list Libosinfo@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libosinfo
