[GRASS-dev] [GRASS GIS] #312: wxNVIZ prints debug to terminal

GRASS GIS trac at osgeo.org
Thu Apr 9 12:04:05 PDT 2015


#312: wxNVIZ prints debug to terminal
-----------------------+----------------------------------------------------
 Reporter:  msieczka   |       Owner:  grass-dev@…              
     Type:  defect     |      Status:  new                      
 Priority:  minor      |   Milestone:  6.4.3                    
Component:  wxGUI      |     Version:  svn-releasebranch64      
 Keywords:  digitizer  |    Platform:  All                      
      Cpu:  All        |  
-----------------------+----------------------------------------------------

Comment(by glynn):

 Replying to [comment:4 wenzeslaus]:

 > > Alternatively, identify the G_message() calls and change them to
 G_verbose_message() so that they don't get printed without --verbose.
 >
 > I don't understand the issue very well. Does this mean that we need to
 change all `G_message()` calls in the library if there is a chance that
 the part will be used directly by some GUI application (vdigit, nviz,
 iclass)?

 First, should GUI programs be different to non-GUI programs in this
 regard?

 The issue is that library functions are generating messages, meaning that
 any program which uses those functions will generate messages. It's
 debatable whether library functions have enough context to decide which
 messages are appropriate.

 Vlib is particularly noisy in this regard. I count 113 calls to
 G_message() in lib/*/*.c (excluding test suites). 59 of those (more than
 half) are in Vlib.

 The other issue seems to be that GUI-format messages are being written to
 the terminal, which suggests an issue in the way that the program is
 started (i.e. something is executing it with GRASS_MESSAGE_FORMAT=gui but
 not capturing its output; the point of GRASS_MESSAGE_FORMAT=gui is to use
 a machine-readable format so that a GUI can process the messages).

-- 
Ticket URL: <http://trac.osgeo.org/grass/ticket/312#comment:5>
GRASS GIS <http://grass.osgeo.org>



More information about the grass-dev mailing list