Another, probably better, way to do it would be to define a new flag like XkbSA_HasGroupFlags inside the XkbModAction flags field when group_flags and group_XXX are valid rather than potentially garbage. That would avoid the whole version-negotiation nightmare, as nothing appears to be too picky about extra flags being defined.
Five years later, it would also be good to have support inside libxkbcommon (which has a pretty decent test suite) and xcb-proto for the flags.
(In reply to Andreas Wettstein from comment #190) /lists. x.org/archives/ xorg-devel/ 2012-November/ 034427. html /lists. x.org/archives/ xorg-devel/ 2012-November/ 034430. html /lists. x.org/archives/ xorg-devel/ 2012-November/ 034429. html /lists. x.org/archives/ xorg-devel/ 2012-November/ 034431. html /lists. x.org/archives/ xorg-devel/ 2012-November/ 034428. html
> No news since. Apart from the formal proposal, there are some old patches
> for its implementation:
> https:/
> https:/
> https:/
> https:/
> https:/
Here's what I think we would need to do in order to not break old clients: /lists. freedesktop. org/archives/ xorg-devel/ 2013-January/ 035049. html
https:/
Another, probably better, way to do it would be to define a new flag like XkbSA_HasGroupFlags inside the XkbModAction flags field when group_flags and group_XXX are valid rather than potentially garbage. That would avoid the whole version-negotiation nightmare, as nothing appears to be too picky about extra flags being defined.
Five years later, it would also be good to have support inside libxkbcommon (which has a pretty decent test suite) and xcb-proto for the flags.