[GRASS-dev] [GRASS GIS] #2846: g.gui.rlisetup: show where new config file is created

[GRASS GIS]

#2846: g.gui.rlisetup: show where new config file is created
--------------------------+---------------------------------
  Reporter:  hellik       |      Owner:  grass-dev@…
      Type:  enhancement  |     Status:  reopened
  Priority:  normal       |  Milestone:  7.4.0
 Component:  Raster       |    Version:  svn-releasebranch70
Resolution:               |   Keywords:
       CPU:  Unspecified  |   Platform:  Unspecified
--------------------------+---------------------------------

Comment (by hellik):

 Replying to [comment:5 mlennert]:
 > Replying to [comment:4 hellik]:
 > > Replying to [comment:3 lucadelu]:
 > > > Replying to [comment:2 mlennert]:
 > > > > Replying to [ticket:2846 hellik]:
 > > > > > it would be nice if there would be an information of the whole
 path to the file which could be copy/paste as input to the r.li.*-modules.
 > > > >
 > > > > You should not use the absolute path in the r.li.* modules. Just
 the file name. You can list existing files with g.gui.rlisetup. So,
 normally, no need to see the absolute path.
 > > > >
 > > >
 > > > I confirm this
 > > >
 > > > > However, ascii output files are also stored in the same path, and
 this is a bit more of a nuisance as you then have to navigate there to
 find the file. I imagine (but haven't checked) that this was done to
 ensure that any other r.li.* module can potentially, easily find the ascii
 output of another, but are there many modules that read this kind of
 output ?
 > > > >
 > > >
 > > > Maybe you can open an enhancement ticket for this?
 > > >
 > > > Can I suggest to close this ticket as invalid? I think it is better
 to use g.gui.rlisetup to create or modify the r.li configuration files.
 > >
 > > What kind of invalid is to show the path to config (and maybe the
 output) files?
 >
 > I think that a message with the full path will give the illusion that
 this is what users should use, while what they actually should use is just
 the name. Maybe to make that even clearer, the r.li.* modules should
 provide a dropdown list of available config files (such as the list of
 signature files you can get in i.maxlik). Again, something for an
 enhancement ticket.
 >

 a dropdown list would be indeed  the best solution.

 > So, maybe invalid is a bit harsh, and wontfix is better, but I agree
 with Luca on the conclusion.

 as long there is no dropdown list, IMHO the user is left alone to know
 names if there are several config files.

--
Ticket URL: <https://trac.osgeo.org/grass/ticket/2846#comment:11>
GRASS GIS <https://grass.osgeo.org>