[wayland HiDPI support, posible regression?]

[microcai]

on Monday 16 March 2015 20:28:48，Daniel Stone wrote：
> Hi,
> 
> On 16 March 2015 at 00:35, Jason Ekstrand <jason at jlekstrand.net> wrote:
> > On Tue, Mar 10, 2015 at 12:00 PM, Bill Spitzak <spitzak at gmail.com> wrote:
> >> [blah blah blah]
> >> 
> >> Events seem to be ok, but my complaint is that a large number of
> >> coordinates in the api other than events are in integer logical pixels,
> >> not in high dpi or in fixed-point. The offsets to attach are the biggest
> >> culprits. There are also integer clip rectangles in the subsurface and
> >> scaling apis. Except for compatibility there is no reason positions in
> >> messages cannot be in buffer pixels.
> 
> 'Except for compatibility', yeah. That's like saying that there's no
> reason for me to have a job, except for the need to house and feed
> myself. Kind of a showstopper, that.
> 
> Smart clients do not require buffer scaling. The scaling is there as a
> fallback to make clients who are blissfully unaware of the constraints
> of high-DPI screens still work: no more, no less. Clients who have the
> smarts to deal with resolution/DPI-independence will _already_ be
> doing smart layout which avoids the need for buffer scaling.

any client *has to be* smart client anyway. buffer scalling is such pointless.


> 
> Any talk of throwing away buffer scaling (breaking dumb clients) in
> order to fit the uses of clients who already today avoid buffer
> scaling, is utterly pointless. Any attempt to handwave away the
> disadvantages as nonexistent is disingenuous.
> 
> > Please do not take a thread started by someone who is obviously
> > confused and side-track it into a discussion of things that you think
> > are design-flaws in the current protocol.  This is not the appropriate
> > place for a discussion of wl_surface.attach (x, y) coordinate systems
> > and bringing that up only adds to the confusion.
> 
> Yes, exactly.
> 
> Yet again, this is something you have repeatedly brought up every time
> something even tangentially related is mentioned. You've explained
> your concerns over and over, and it's obvious that our opinions differ
> and upstream will not change. Doing this makes it infinitely less
> likely that your concerns will be taken seriously (cf. the
> wl_keyboard_grab bug): the first reaction to seeing your name come up
> in a thread is 'oh god, not again'. Which is a shame, as you do have
> valuable input to offer, but it's drowned out by the amount that you
> bang on about your pet peeves, with a total inability to accept that
> someone with a differing opinion may just have a different opinion,
> not be objectively wrong. Everyone loses: you don't get taken
> seriously, we get frustrated, discussions get derailed, and people who
> don't know better mistake your loud pronouncements for upstream's
> actual position (or, when those differ from measurable reality rather
> than opinion, a useful fact).
> 
> Be the signal, not the noise.
> 
> Cheers,
> Daniel