Re: dte: darwin doesn't get X11-specific files

[Jason Bacon <jtocino%gmx.com@localhost>]

On 1/3/24 08:41, Greg Troxel wrote:
> Kevin Bloom <kevin.bloom%posteo.net@localhost> writes:
>
> > This is an error in my commit message. I only used that lingo because
> > normally .desktop and app xml files are used by desktop environments
> > or window managers (i.e. graphical menus). Jason [Bacon] has a
> > different solution of keeping the files in the PLIST and added a
> > patch for the GNUmakefile that forces those files to be installed
> > on darwin. He posted his patch to the pkgsrc-usrs mailing list so you
> > can see it there. He also noticed that I did the PLIST exclusion
> > incorrect.
> >
> > Since I'm the new guy I'm fine with whichever appoarch is "best."
>
> I agree that fixing upstream not to conditionalize is best.  I think
> it's a bug for them to assume that darwin means not x.  I think we've
> landed in a good place.

I haven't been watching wip-changes, but I just read the thread.

XQuartz (X11 server for Mac) works pretty well, but the situation is
less than awesome.  It takes a while to fire up the X11 server
on-demand, and it provides a UI that's inconsistent with the Mac
interface.  There have also been numerous bugs over the years that have
caused some X11 apps not to function.

Institutionally-managed Macs may not have XQuartz installed and it can
be a headache to request it.  I.T. departments are defensive by nature,
and hesitant to roll out software they're not familiar with.  Those
fears are often well-justified, as it takes a lot of resources to deal
with surprises.

It's probably best to assume that Mac users don't have X11.  If I were
developing cross-platform GUI apps that I wanted to work on Mac, I would
use Qt so that X11 isn't necessary, and the Mac UI is more standard.

I would only disable X11 features that cause real problems, though.
Better to leave other stuff in for users who do have X11 and know how to
use it.  The DTE Darwin check is overaggressive in my view.

Cheers,

	J