Re: Connecting to a modern Xserver with an old client
On Mon, 24 Jul 2023 20:56:30 +1000 Peter Hutterer said: > On Mon, Jul 24, 2023 at 10:56:13AM +0100, Dave Howorth wrote: > > On Sun, 23 Jul 2023 14:14:37 -0700 > > Alan Coopersmith wrote: > > > > > On 7/23/23 13:48, Carsten Haitzler wrote: > > > >> So as per the man page I added +byteswappedclients to the Xorg > > > >> command line and it works! > > > > > > > > wow... i've never encountered that before. in all these decades... > > > > well well. :) > > > > > > It was only added a few months ago: > > > https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1029 > > > > > > Xwayland 23.1 was the first stable release to ship it - on the Xorg > > > side I don't think it's in any stable release yet, just the git repo. > > > > It sounds like there needs to be a clearer error message produced, > > given the amount of hassle it's taken to identify this problem. I'm not > > sure whether the server or the client ought to report it explicitly, > > but I guess it would be a smaller change to ensure that a server with > > the new feature logged any failures that it caught/caused? > > I'm not really disagreeing but we don't have a lot of options here, > doubly so for old clients that won't be modified. We do use the proper > client error reporting so ideally you'll see that on the terminal - alas > if you're not starting the client from the terminal you won't see that. > > And logging every access from a client is not great either - and logging > it once will likely just make it disappear in the rest of the server log. I think the suggestion to just log this specific error (endianess) and only this would be a rare event but worth logging on the server side as at least you'd have something to go by. it'd otherwise be as quiet as it is now. -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com
Re: X (Twitter) Logo
On Mon, 24 Jul 2023 07:52:21 -0700 Jeremy Sequoia wrote: > Did Elon just steal our logo for Twitter's rebranding? No, he stole the "blackboard bold" X from Unicode: http://www.unicode-symbol.com/u/1D54F.html Attila Kinali -- In science if you know what you are doing you should not be doing it. In engineering if you do not know what you are doing you should not be doing it. -- Richard W. Hamming, The Art of Doing Science and Engineering
Re: Connecting to a modern Xserver with an old client
On Sun, 23 Jul 2023 14:14:37 -0700 Alan Coopersmith said: > On 7/23/23 13:48, Carsten Haitzler wrote: > >> So as per the man page I added +byteswappedclients to the Xorg command line > >> and it works! > > > > wow... i've never encountered that before. in all these decades... well > > well. :) > > It was only added a few months ago: > https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1029 It seems we got our first "victim" of that new change. :) > Xwayland 23.1 was the first stable release to ship it - on the Xorg side > I don't think it's in any stable release yet, just the git repo. > > -- > -Alan Coopersmith- alan.coopersm...@oracle.com > Oracle Solaris Engineering - https://blogs.oracle.com/solaris > -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com
X (Twitter) Logo
Did Elon just steal our logo for Twitter's rebranding?
Re: Connecting to a modern Xserver with an old client
On Mon, Jul 24, 2023 at 12:56 PM Peter Hutterer wrote: > And logging every access from a client is not great either - and logging > it once will likely just make it disappear in the rest of the server log. Well, you just need to log _failing_ clients. Those clients will not produce many requests as they are rejected right at the beginning, not getting a display. Right? But just to mention it: I think rejecting them is not the right way to handle them anyway. You never know what clients you'll see in the future and having to restart X to give them access is a no-go in my POV. Uli
Re: Connecting to a modern Xserver with an old client
On Mon, Jul 24, 2023 at 10:56:13AM +0100, Dave Howorth wrote: > On Sun, 23 Jul 2023 14:14:37 -0700 > Alan Coopersmith wrote: > > > On 7/23/23 13:48, Carsten Haitzler wrote: > > >> So as per the man page I added +byteswappedclients to the Xorg > > >> command line and it works! > > > > > > wow... i've never encountered that before. in all these decades... > > > well well. :) > > > > It was only added a few months ago: > > https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1029 > > > > Xwayland 23.1 was the first stable release to ship it - on the Xorg > > side I don't think it's in any stable release yet, just the git repo. > > It sounds like there needs to be a clearer error message produced, > given the amount of hassle it's taken to identify this problem. I'm not > sure whether the server or the client ought to report it explicitly, > but I guess it would be a smaller change to ensure that a server with > the new feature logged any failures that it caught/caused? I'm not really disagreeing but we don't have a lot of options here, doubly so for old clients that won't be modified. We do use the proper client error reporting so ideally you'll see that on the terminal - alas if you're not starting the client from the terminal you won't see that. And logging every access from a client is not great either - and logging it once will likely just make it disappear in the rest of the server log. Cheers, Peter
Re: Connecting to a modern Xserver with an old client
On Sun, 23 Jul 2023 14:14:37 -0700 Alan Coopersmith wrote: > On 7/23/23 13:48, Carsten Haitzler wrote: > >> So as per the man page I added +byteswappedclients to the Xorg > >> command line and it works! > > > > wow... i've never encountered that before. in all these decades... > > well well. :) > > It was only added a few months ago: > https://gitlab.freedesktop.org/xorg/xserver/-/merge_requests/1029 > > Xwayland 23.1 was the first stable release to ship it - on the Xorg > side I don't think it's in any stable release yet, just the git repo. It sounds like there needs to be a clearer error message produced, given the amount of hassle it's taken to identify this problem. I'm not sure whether the server or the client ought to report it explicitly, but I guess it would be a smaller change to ensure that a server with the new feature logged any failures that it caught/caused?