[GRASS-user] Trouble getting started

GRASS GRASS user
1

I am totally new to grass and having trouble.

I installed grass from RPM on my Fedora 30 linux system. First I installed 7.6 from official fedora packages, then I erased that and installed 7.8 from packages on the Fedora "copr" repository. I downloaded the sample North Carolina dataset for version 7 (145M). I untarred that into /home/tom/grassdata and when I browse to that, grass seems to find it and is ready to go. But when I click on "Start Grass Session" I get two empty and useless windows and see the messages:

(wxgui.py:30767): Gtk-CRITICAL **: 20:37:16.297: gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkScrollbar

(wxgui.py:30767): Gtk-CRITICAL **: 20:37:16.297: gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkScrollbar

I get the exact same with 7.6 and 7.8.

The useless windows call themselves "layer manager" and "map display".

This seems like a reasonable way to get started. What if anything am I doing wrong, or is this some known bug?

2

Hi Tom,

Le Thu, 14 Nov 2019 20:40:31 -0700,
Tom Trebisky <tom@mmto.org> a écrit :

I am totally new to grass and having trouble.

Welcome to the community !

GRASS GIS does have a bit of a learning curve, but once you get past
the first barriers you won't be able to live without it anymore ! :wink:

I installed grass from RPM on my Fedora 30 linux system. First I
installed 7.6 from official fedora packages, then I erased that and
installed 7.8 from packages on the Fedora "copr" repository. I
downloaded the sample North Carolina dataset for version 7 (145M). I
untarred that into /home/tom/grassdata and when I browse to that,
grass seems to find it and is ready to go. But when I click on
"Start Grass Session" I get two empty and useless windows

When you say "empty", what does that mean exactly ? Could you provide a
screenshot ?

and see the
messages:

(wxgui.py:30767): Gtk-CRITICAL **: 20:37:16.297:
gtk_box_gadget_distribute: assertion 'size >= 0' failed in
GtkScrollbar

(wxgui.py:30767): Gtk-CRITICAL **: 20:37:16.297:
gtk_box_gadget_distribute: assertion 'size >= 0' failed in
GtkScrollbar

These are just annoying messages, but harmless and AFAIK do not
explain your issue.

I get the exact same with 7.6 and 7.8.

The useless windows call themselves "layer manager" and "map display".

These two windows should open by default when you launch the GRASS GIS
GUI, but they should contain buttons and menus.

If they do, then the next step for you should be to use the layer
manager window to add a map which should then appear in the map display
window.

Have you had the opportunity to read these to introductions:

https://grass.osgeo.org/grass78/manuals/helptext.html
https://grasswiki.osgeo.org/wiki/Quick_wxGUI_tutorial

Also check here:

https://grass.osgeo.org/documentation/tutorials/

for some more in depth tutorials.

This seems like a reasonable way to get started. What if anything am
I doing wrong, or is this some known bug?

At this stage we need a bit more info to really know.

Moritz

3

Le Fri, 15 Nov 2019 11:27:54 +0100,
Moritz Lennert <mlennert@club.worldonline.be> a écrit :

Hi Tom,

Le Thu, 14 Nov 2019 20:40:31 -0700,
Tom Trebisky <tom@mmto.org> a écrit :

> I am totally new to grass and having trouble.

Welcome to the community !

GRASS GIS does have a bit of a learning curve, but once you get past
the first barriers you won't be able to live without it anymore ! :wink:

>
> I installed grass from RPM on my Fedora 30 linux system. First I
> installed 7.6 from official fedora packages, then I erased that and
> installed 7.8 from packages on the Fedora "copr" repository.

Another question: when exactly did you install these packages. 7.8.1
just arrived in the repository 3 days ago, IISC, so if you still have
7.8.0 you might want to start by installing the latest version.

Moritz

4

Hi Moritz,

Thank you for your help and advice. I am running 7.8.1RC1 as it turns out, installed yesterday.

I was overly alarmed by the error messages (they do say "Critical Error" after all) and along with
getting two blank windows, I just assumed that something serious was wrong.

I do think that it would make sense for somebody to fix whatever it is that yields the messages like:

(wxgui.py:419): Gtk-CRITICAL **: 21:21:42.695: gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkSpinButton

In fact I note that the window I used to start GRASS is filled with messages of this sort.
Likely that the Gtk libraries have become extra fussy or something, but it would be nice to clean this
up as it will alarm new users (or any sensible user I would think).

My blank windows do have buttons and when I ignore the error messages that continue to spew out, I do get
results! I had expected that I would immediately see some kind of map, not just a blank canvas to start, so on
with the learning curve I go.

 Tom

On 11/15/19 3:58 AM, Moritz Lennert wrote:

Le Fri, 15 Nov 2019 11:27:54 +0100,
Moritz Lennert <mlennert@club.worldonline.be> a écrit :

Hi Tom,

Le Thu, 14 Nov 2019 20:40:31 -0700,
Tom Trebisky <tom@mmto.org> a écrit :

I am totally new to grass and having trouble.

Welcome to the community !

GRASS GIS does have a bit of a learning curve, but once you get past
the first barriers you won't be able to live without it anymore ! :wink:

I installed grass from RPM on my Fedora 30 linux system. First I
installed 7.6 from official fedora packages, then I erased that and
installed 7.8 from packages on the Fedora "copr" repository.

Another question: when exactly did you install these packages. 7.8.1
just arrived in the repository 3 days ago, IISC, so if you still have
7.8.0 you might want to start by installing the latest version.

Moritz

5

Hi Tom,

On Fri, Nov 15, 2019 at 5:05 PM Tom Trebisky <tom@mmto.org> wrote:

Hi Moritz,

Thank you for your help and advice. I am running 7.8.1RC1 as it turns
out, installed yesterday.

I was overly alarmed by the error messages (they do say "Critical Error"
after all) and along with
getting two blank windows, I just assumed that something serious was wrong.

Yes, these GTK error messages are annoying a big time!
Seems nobody knows how to get rid of them. See your preferred search
engine that many software packages with terminal support + GUI suffer
from them...

I do think that it would make sense for somebody to fix whatever it is
that yields the messages like:

(wxgui.py:419): Gtk-CRITICAL **: 21:21:42.695:
gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkSpinButton

Fully agreed. The thing is that they are rather impossible to debug.
It says line 419 but there is nothing related.

In fact I note that the window I used to start GRASS is filled with
messages of this sort.
Likely that the Gtk libraries have become extra fussy or something, but
it would be nice to clean this
up as it will alarm new users (or any sensible user I would think).

Again, fully agreed.

My blank windows do have buttons and when I ignore the error messages
that continue to spew out, I do get
results! I had expected that I would immediately see some kind of map,
not just a blank canvas to start, so on
with the learning curve I go.

That's a different topic but also agreed :wink:

Markus

6

I see. Well I have done my share of Gtk programming in C and via Ruby and I remember things of this sort (messages that were difficult or impossible to relate to code), especially when callbacks are registered. I presume (or get the impression) that you are a developer and not just an end user of Grass.

I find myself half tempted to clone the Git repository and try to build Grass from sources, but this is not why I got involved here. And I am sure better minds than mine have already looked at this issue.

I might presume/guess that this has been identified as an upstream issue in the Python/Gtk bindings or perhaps even in Gtk itself -- and there may be better things to work on than chasing this in hopes that the teams upstream will do something about all of this.

My interests (with regard to Grass) are in maps and geology and I will try to stick to that for now.

Thanks for your friendly reception into the Grass community !!!

 Tom

On 11/15/19 3:26 PM, Markus Neteler wrote:

Hi Tom,

On Fri, Nov 15, 2019 at 5:05 PM Tom Trebisky <tom@mmto.org> wrote:

Hi Moritz,

Thank you for your help and advice. I am running 7.8.1RC1 as it turns
out, installed yesterday.

I was overly alarmed by the error messages (they do say "Critical Error"
after all) and along with
getting two blank windows, I just assumed that something serious was wrong.

Yes, these GTK error messages are annoying a big time!
Seems nobody knows how to get rid of them. See your preferred search
engine that many software packages with terminal support + GUI suffer
from them...

I do think that it would make sense for somebody to fix whatever it is
that yields the messages like:

(wxgui.py:419): Gtk-CRITICAL **: 21:21:42.695:
gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkSpinButton

Fully agreed. The thing is that they are rather impossible to debug.
It says line 419 but there is nothing related.

In fact I note that the window I used to start GRASS is filled with
messages of this sort.
Likely that the Gtk libraries have become extra fussy or something, but
it would be nice to clean this
up as it will alarm new users (or any sensible user I would think).

Again, fully agreed.

My blank windows do have buttons and when I ignore the error messages
that continue to spew out, I do get
results! I had expected that I would immediately see some kind of map,
not just a blank canvas to start, so on
with the learning curve I go.

That's a different topic but also agreed :wink:

Markus

7

On 16/11/19 03:10, Tom Trebisky wrote:

I see. Well I have done my share of Gtk programming in C and via Ruby and I remember things of this sort (messages that were difficult or impossible to relate to code), especially when callbacks are registered. I presume (or get the impression) that you are a developer and not just an end user of Grass.

I find myself half tempted to clone the Git repository and try to build Grass from sources, but this is not why I got involved here. And I am sure better minds than mine have already looked at this issue.

I might presume/guess that this has been identified as an upstream issue in the Python/Gtk bindings or perhaps even in Gtk itself -- and there may be better things to work on than chasing this in hopes that the teams upstream will do something about all of this.

See https://trac.osgeo.org/grass/ticket/3348.

If you have any suggestions, they are obviously welcome. :slight_smile:

Moritz