On 09/27/22 17:59, Daniel P. Berrangé wrote: > On Tue, Sep 27, 2022 at 04:53:08PM +0200, Laszlo Ersek wrote: >> On 09/26/22 14:09, Daniel P. Berrangé wrote: >>> On Mon, Sep 26, 2022 at 12:59:21PM +0100, Richard W.M. Jones wrote: >>>> >>>> Apart from Dan's suggestions in patch 1, the series looks good to me. >>>> >>>> FWIW RHEL 7 (the earliest distro with PCRE 2) has glib2 2.56.1 & >>>> gtk3 3.22.30, so supporting any earlier versions also seems pointless, >>>> so that might be another thing to review. We could make USE_POPOVERS >>>> unconditional, and make gui-gtk3-compat.h considerably less >>>> complicated. >>> >>> Say you pick 3.22 as your official min and check this with pkg-config >>> in configure, then you can further define >>> >>> GDK_VERSION_MIN_REQUIRED=GDK_VERSION_3_22 >>> GDK_VERSION_MAX_ALLOWED=GDK_VERSION_3_22 >>> >>> The former will give you warnings if you use any API that was already >>> deprecated in 3.22 - this is something that should be re-written to >>> use the recommended modern alternative API. >> >> Is "GDK" a muscle-memory typo (because "GDK" does exist)? Did you mean GTK? > > It looks odd, but I really do mean GDK here. It affects both the GDK > and GTK include files (GDK is a layer below GTK).
Right, I've found a diagram here: https://en.wikipedia.org/wiki/GDK What I didn't know was that the GTK and GDK versions were supposed to be in sync. In virt-p2v, we care about GTK (and maybe glib); I've not been aware that we care about APIs at the GDK level. So enforcing a GDK version seemed strange, as a proxy for the GTK version. Laszlo _______________________________________________ Libguestfs mailing list Libguestfs@redhat.com https://listman.redhat.com/mailman/listinfo/libguestfs
