honour window movements

[totaam]

Issue migrated from trac ticket # 212

component: core | priority: minor | resolution: fixed

2012-11-13 09:19:27: antoine created the issue

---

since r1089 we honour client triggered window resizes (see #107), we should probably also honour client window movements.
Unfortunately this is more tricky... see #158

Problems:

• avoid loops, the client will send a configure event when the window is moved... which can trigger another move, etc..
• as per preserve workspace mapping #170: there are small offsets to do with window borders - this needs sorting out too
• too many places where the same information is kept, making it difficult to synchronize everything - maybe more of it should be derived/calculated rather than stored?

[totaam]

2012-11-13 09:25:00: antoine uploaded file honour-window-move.patch (3.6 KiB)

if we find that the client window has moved, move our corral window and pass it on to the client

[totaam]

2012-11-13 09:27:16: antoine changed status from new to accepted

[totaam]

2012-11-13 09:27:16: antoine edited the issue description

[totaam]

2012-11-13 09:28:38: antoine uploaded file honour-window-movep2.patch (4.0 KiB)

missing bits from the first patch

[totaam]

2013-03-20 14:50:25: antoine commented

---

re-scheduling for a later version

[totaam]

2013-10-17 12:35:26: totaam commented

---

0.12 should deal with focus/move issues

[totaam]

2014-03-16 11:51:40: totaam changed status from accepted to new

[totaam]

2014-03-16 11:51:40: totaam commented

---

Many changes in this area because of #532, and window moves now done in r5811.

Please check on a variety of platforms using the test applications:

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_move.py test_window_move]
• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_resize.py test_window_resize]
• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_mapresize.py test_window_mapresize]

Especially with tricksy setups like multi-monitor: when the windows are on the secondary one (to make sure the coordinates used for the calculations are absolute and not relative to whatever screen you are on)

And maybe also check for regressions in fullscreen/maximize/etc (just in case):

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_maximize.py test_window_maximize]
  etc.

[totaam]

2014-03-16 11:51:49: totaam changed owner from antoine to afarr

[totaam]

2014-03-19 01:03:22: afarr commented

---

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_move.py test_window_move] - on osx this seems to work as expected, but on windows the move me button doesn't move the window - though if moved manually, the location is output in the xterm from which the process is spawned.

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_resize.py test_window_resize] - seems to work as expected, osx and windows (though the fast resize is not really very fast at all on osx).

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_mapresize.py test_window_mapresize] - on both osx and windows it seems to work as expected, but after ~ 8 seconds the window closes (whether any resizing has been done or not) ... coordinates do not seem to be relative to the screen, but rather relative to the total display size.

• [/browser/xpra/trunk/src/tests/xpra/test_apps/test_window_maximize.py test_window_maximize] - this seems to work as expected, on a single window osx or multi-window windows set up, though the unmaximize me button doesn't work, which is easily worked around by using the unmaximize button supplied by the OS.

[totaam]

2014-03-19 06:16:26: totaam commented

---

• move: I am unable to reproduce the problem on win32, what revision are you testing with? can you post the client and server log with "-d window" (include only the relevant section)
• resize: fast_resize may well behave differently on different platforms, depending on how quickly they respond to our request to resize the window (more slowly on OSX it seems). As long as things end up looking OK, we're good.
• mapresize: it was meant to close after 10 seconds, I've changed it in r5851
• maximize: that's a bug: handle unmaximize #537

[totaam]

2014-03-20 21:30:12: afarr commented

---

I think the version was the issue with the test_window_move - I was using r5444.

Testing again with r5828 the move me button causes the window to "hop" as expected.

This looks like it should all be good. (I assume you no longer need the client and server logs?)

[totaam]

2014-03-21 00:37:54: totaam changed status from new to closed

[totaam]

2014-03-21 00:37:54: totaam changed resolution from ** to fixed

[totaam]

2014-03-21 00:37:54: totaam commented

---

Done!