Close wxGUI on GRASS CLI exit

[marisn]

When user exits from GRASS CLI (terminal), Init.sh should issue a close command to any remaining wxGUI processes.

See r39350 for ideas how it's done in gis.m.

Reported by Martin: ​http://lists.osgeo.org/pipermail/grass-dev/2009-September/046460.html

[cmbarton]

This needs to be done in init.sh I believe.

Michael

[neteler]

Still to be done (both G6 and G7).

[mlennert]

I wouldn't consider this a bug. Reclassifying as enhancement wish.

[wenzeslaus]

Please see and try trunk after r66917, r66918, r66920, and r66982 by martinl.

[martinl]

Replying to wenzeslaus:

Please see and try trunk after r66917, r66918, r66920, and r66982 by martinl.

right, I didn't announce yet. I was able to test on Windows 7 (cmd), Debian unstable (bash). Any testing is highly welcome! I would like to backport it to relbr70 before 7.0.3RC1 (January 2016).

I also changed WinGRASS (currently only trunk, standalone installator) to start cmd minimized. I am not sure if it's good idea, it take some time when startup screen appears and user could have feeling that nothing happend (cmd is started minimized) and start GRASS several times. Starting cmd in normal mode is probably better idea in the sense that user has some information that GRASS is starting. On the other side minimized cmd is not so scary for most of Windows users ;-) What do you think?

[martinl]

Replying to wenzeslaus:

Please see and try trunk after r66917, r66918, r66920, and r66982 by martinl.

BTW, these commits enable to close CLI terminal (cmd) automatically when quiting GUI. The ticket reports other issue. To close GUI automaticaly (currently GUI just hangs and become unfunctional) when closing CLI terminal.

[hellik]

Replying to martinl:

Starting cmd in normal mode is probably better idea in the sense that user has some information that GRASS is starting. On the other side minimized cmd is not so scary for most of Windows users ;-) What do you think?

Launching <wxpython> GUI in the background, please wait...

this line is the important information already in cmd. I may prefer the the normal mode. windows users should be scared of windows, not of grass ... ;-)

[mlennert]

Replying to martinl:

Replying to wenzeslaus:

Please see and try trunk after r66917, r66918, r66920, and r66982 by martinl.

right, I didn't announce yet. I was able to test on Windows 7 (cmd), Debian unstable (bash). Any testing is highly welcome! I would like to backport it to relbr70 before 7.0.3RC1 (January 2016).

I also changed WinGRASS (currently only trunk, standalone installator) to start cmd minimized. I am not sure if it's good idea, it take some time when startup screen appears and user could have feeling that nothing happend (cmd is started minimized) and start GRASS several times. Starting cmd in normal mode is probably better idea in the sense that user has some information that GRASS is starting. On the other side minimized cmd is not so scary for most of Windows users ;-) What do you think?

-1 (to minimizing cmd)

I find it useful to have cmd visible, including for teaching.

Moritz

[martinl]

Replying to mlennert:

I also changed WinGRASS (currently only trunk, standalone installator) to start cmd minimized. I am not sure if it's good idea, it take some time when startup screen appears and user could have feeling that nothing happend (cmd is started minimized) and start GRASS several times. Starting cmd in normal mode is probably better idea in the sense that user has some information that GRASS is starting. On the other side minimized cmd is not so scary for most of Windows users ;-) What do you think?

-1 (to minimizing cmd)

I find it useful to have cmd visible, including for teaching.

OK, reverted in r67119

[wenzeslaus]

Replying to martinl:

Replying to wenzeslaus:

Please see and try trunk after r66917, r66918, r66920, and r66982 by martinl.

BTW, these commits enable to close CLI terminal (cmd) automatically when quiting GUI. The ticket reports other issue. To close GUI automaticaly (currently GUI just hangs and become unfunctional) when closing CLI terminal.

The GUIs left behind after closing GRASS GIS now seem to be a bigger problem. On thing is that from user perspective the command line and GUI are now more connected. (Fortunately, after r66920, there is no more the danger of terminating a different process.) BTW, I like it much more after r67244.

[martinl]

Replying to wenzeslaus:

The GUIs left behind after closing GRASS GIS now seem to be a bigger problem. On thing is that from user perspective the command line and GUI are now more connected. (Fortunately, after

attempt to solve this issue applied in r67306

[martinl]

Replying to martinl:

attempt to solve this issue applied in r67306

backported to relbr70 in r67731. Please test.

[martinl]

Tested closing. Feel free to reopen if needed.