Hi,
I was running sz81, a ZX81 emulator using SDL, remotely with an IRC client and noticed, that while it worked nicely if started in a session using an x2go Linux client, after disconnecting the session and reconnecting using a Windows client, the application window came up garbled with display artifacts.
What's more, if I disconnect and reconnect the session using a Linux client, it somehow manages to blank my complete local display as soon as the SDL application window gets focus. This confused the hell out of me, and I first suspected a crash, until I figured out what was going on. It's really an SDL issue, I verified by running parallax-4.tar.gz from http://olofson.net/examples.html which displayed the same behaviour.
I would guess x2go has some kind of SDL passthrough mode which is somewhat buggy and doesn't work with windows. Is that correct, or what is the expected behaviour?
Hi,
I don't know anything about SDL so all I can do is some speculation. Maybe SDL uses some X11 operations that are not handled correctly.
What version of x2goagent are you using on the server side (x2goagent -version)?
Uli
On Fri, Feb 10, 2023 at 10:56 PM Hanno Foest <hurga-x2go@tigress.com> wrote:
Hi,
I was running sz81, a ZX81 emulator using SDL, remotely with an IRC client and noticed, that while it worked nicely if started in a session using an x2go Linux client, after disconnecting the session and reconnecting using a Windows client, the application window came up garbled with display artifacts.
What's more, if I disconnect and reconnect the session using a Linux client, it somehow manages to blank my complete local display as soon as the SDL application window gets focus. This confused the hell out of me, and I first suspected a crash, until I figured out what was going on. It's really an SDL issue, I verified by running parallax-4.tar.gz from http://olofson.net/examples.html which displayed the same behaviour.
I would guess x2go has some kind of SDL passthrough mode which is somewhat buggy and doesn't work with windows. Is that correct, or what is the expected behaviour?
- Hanno
x2go-user mailing list x2go-user@lists.x2go.org https://lists.x2go.org/listinfo/x2go-user
I played around a bit: I could only compile parallax-2, the newer ones show a linking problem: gcc -g -O2 -I/usr/include/SDL -D_GNU_SOURCE=1 -D_REENTRANT -o parallax3 parallax3.o -L/usr/lib/x86_64-linux-gnu -lSDL /usr/bin/ld: parallax3.o: undefined reference to symbol 'cos@@GLIBC_2.2.5' /usr/bin/ld: /lib/x86_64-linux-gnu/libm.so.6: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status make: *** [Makefile:157: parallax3] Fehler 1 Nevertheless I made a quick test: 1. run nxagent -ac :55 2. in a second shell run DISPLAY=:55 parallax2 -> works fine on the nxagent window 3. in a third shell run kill -HUP $(pgrep nxagent) -> window will close (resembling session suspend) 4. wait a while, then repeat the previous command -> window will show again (resembling session resume) -> no fullscreen, no glitches Repeating the test with nxagent -ac -55 -R in step 1 (thus simulating a rootless session) shows the problem here, too. So it really looks like a bug in nxagent. Can you confirm? Uli On Sat, Feb 11, 2023 at 3:06 PM Ulrich Sibiller <uli42@gmx.de> wrote:
Hi,
I don't know anything about SDL so all I can do is some speculation. Maybe SDL uses some X11 operations that are not handled correctly.
What version of x2goagent are you using on the server side (x2goagent -version)?
Uli
On Fri, Feb 10, 2023 at 10:56 PM Hanno Foest <hurga-x2go@tigress.com> wrote:
Hi,
I was running sz81, a ZX81 emulator using SDL, remotely with an IRC client and noticed, that while it worked nicely if started in a session using an x2go Linux client, after disconnecting the session and reconnecting using a Windows client, the application window came up garbled with display artifacts.
What's more, if I disconnect and reconnect the session using a Linux client, it somehow manages to blank my complete local display as soon as the SDL application window gets focus. This confused the hell out of me, and I first suspected a crash, until I figured out what was going on. It's really an SDL issue, I verified by running parallax-4.tar.gz from http://olofson.net/examples.html which displayed the same behaviour.
I would guess x2go has some kind of SDL passthrough mode which is somewhat buggy and doesn't work with windows. Is that correct, or what is the expected behaviour?
- Hanno
x2go-user mailing list x2go-user@lists.x2go.org https://lists.x2go.org/listinfo/x2go-user
On Sat, Feb 11, 2023 at 03:23:49PM +0100, Ulrich Sibiller wrote:
I played around a bit: I could only compile parallax-2, the newer ones show a linking problem:
parallax-4 worked for me with
parallax4_LDFLAGS = "-Wl,--copy-dt-needed-entries"
in Makefile.
Nevertheless I made a quick test:
- run nxagent -ac :55
- in a second shell run DISPLAY=:55 parallax2 -> works fine on the nxagent window
- in a third shell run kill -HUP $(pgrep nxagent) -> window will close (resembling session suspend)
- wait a while, then repeat the previous command -> window will show again (resembling session resume) -> no fullscreen, no glitches
Repeating the test with nxagent -ac -55 -R in step 1 (thus simulating a rootless session) shows the problem here, too. So it really looks like a bug in nxagent.
Can you confirm?
Yes, exactly the same here!
Hanno
I have opened a ticket for this issue: https://github.com/ArcticaProject/nx-libs/issues/1053
On Sat, Feb 11, 2023 at 4:35 PM Hanno Foest <hurga-x2go@tigress.com> wrote:
On Sat, Feb 11, 2023 at 03:23:49PM +0100, Ulrich Sibiller wrote:
I played around a bit: I could only compile parallax-2, the newer ones show a linking problem:
parallax-4 worked for me with
parallax4_LDFLAGS = "-Wl,--copy-dt-needed-entries"
in Makefile.
Nevertheless I made a quick test:
- run nxagent -ac :55
- in a second shell run DISPLAY=:55 parallax2 -> works fine on the nxagent window
- in a third shell run kill -HUP $(pgrep nxagent) -> window will close (resembling session suspend)
- wait a while, then repeat the previous command -> window will show again (resembling session resume) -> no fullscreen, no glitches
Repeating the test with nxagent -ac -55 -R in step 1 (thus simulating a rootless session) shows the problem here, too. So it really looks like a bug in nxagent.
Can you confirm?
Yes, exactly the same here!
Hanno
x2go-user mailing list x2go-user@lists.x2go.org https://lists.x2go.org/listinfo/x2go-user