[GRASS-dev] Re: GUI support for TclTk 8.5

GRASS GIS GRASS GIS developer
1

On Mar 8, 2008, at 1:38 AM, grass-dev-request@lists.osgeo.org wrote:

Date: Fri, 7 Mar 2008 16:06:13 -0800 (PST)
From: Hamish <hamish_b@yahoo.com>
Subject: Re: [GRASS-dev] Current state of GUIs on Win32
To: grass-dev@lists.osgeo.org
Message-ID: <714751.56317.qm@web45803.mail.sp1.yahoo.com>
Content-Type: text/plain; charset=iso-8859-1

Michael Barton wrote:

The current GUI is based on TclTk 8.4.14 or so. As some people have
noted, it may work with 8.5, but there is no guarantee about it.
Given that we are replacing this GUI, I don't think it is worth the
effort to update the GUI to 8.5.

IMO we should at least attempt to support Tcl/Tk 8.3 - 8.5.

For 8.3 there is just the one problem with v.digit, AFAIK.

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

2

Michael:

You are right that fixing issues is probably pretty simple. It's
figuring out what to fix that is time consuming. Right now, I'm
heavily overcommitted timewise. And on top of that, I'm using my
laptop (primary development box) for my GIS class and can't afford to
have GRASS broken for more than a couple days. So I have no problem
in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ
internals or something like that, but won't have time to do much for
awhile.

..... the responsibility does not have to to rest solely on your
shoulders of course; all are encouraged to help test and diagnose
where+what the problematic areas (if any!) are and help document/fix
them.

Hamish

      ____________________________________________________________________________________
Never miss a thing. Make Yahoo your home page.
http://www.yahoo.com/r/hs

3

On Mar 8, 2008, at 5:26 PM, Hamish wrote:

Michael:

You are right that fixing issues is probably pretty simple. It's
figuring out what to fix that is time consuming. Right now, I'm
heavily overcommitted timewise. And on top of that, I'm using my
laptop (primary development box) for my GIS class and can't afford to
have GRASS broken for more than a couple days. So I have no problem
in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ
internals or something like that, but won't have time to do much for
awhile.

..... the responsibility does not have to to rest solely on your
shoulders of course; all are encouraged to help test and diagnose
where+what the problematic areas (if any!) are and help document/fix
them.

Hamish

Definitely much appreciated. You've been especially helpful in this regard.

Michael

4

Hi,

re. r.in.wms XML paring code for layers with spaces in the name

given some text like this:

DATA="<Name>Foo Bar Baz</Name>"
echo "$DATA" | sed -e "s/<Name>\s*\(\w*\)/~\1~/g" -e "s/<\/Name>//g"

you get
~Foo~ Bar Baz

instead of
~Foo Bar Baz~

how to fix that regex?

thanks,
Hamish

      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now. http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

5

Rewrite this in python ?

or using new gdal features

jachym

Hamish píše v So 08. 03. 2008 v 20:27 -0800:

Hi,

re. r.in.wms XML paring code for layers with spaces in the name

given some text like this:

DATA="<Name>Foo Bar Baz</Name>"
echo "$DATA" | sed -e "s/<Name>\s*\(\w*\)/~\1~/g" -e "s/<\/Name>//g"

you get
~Foo~ Bar Baz

instead of
~Foo Bar Baz~

how to fix that regex?

thanks,
Hamish

      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now. http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

--
Jachym Cepicky
e-mail: jachym.cepicky gmail com
URL: http://les-ejk.cz
GPG: http://www.les-ejk.cz/pgp/jachym_cepicky-gpg.pub

6

I can build an OSX GRASS app with TclTk 8.5 for others to test, and try it myself.

If we intend to support multiple versions of TclTk, it would be helpful to have a TCLTKVER variable in Platform.make for the configured major.minor version.

On Mar 8, 2008, at 9:06 AM, Michael Barton wrote:

IMO we should at least attempt to support Tcl/Tk 8.3 - 8.5.

For 8.3 there is just the one problem with v.digit, AFAIK.

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

First Pogril: Why is life like sticking your head in a bucket filled with hyena offal?
Second Pogril: I don't know. Why IS life like sticking your head in a bucket filled with hyena offal?
First Pogril: I don't know either. Wretched, isn't it?

-HitchHiker's Guide to the Galaxy

7

On Mar 8, 2008, at 9:06 AM, Michael Barton wrote:

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

Some initial tests for TclTk 8.5.1 on OSX.

10.4/Tiger: the GUI runs and seems to work properly. NVIZ won't run - with Spearfish:

"
nviz elev=elevation.10m

alloc: invalid block: 0x5202b4: 61 0 0

[2]+ Abort trap
"

Even just "nviz --help" gives that alloc error.

10.5/Leopard: won't run at all. There appears to be some problem between Leopard's fontconfig/freetype and TclTk 8.5 (TclTk 8.4 works).

Thread 0 Crashed:
0 libfontconfig.1.dylib 0x004a620a FcConfigAddCache + 149
1 libfontconfig.1.dylib 0x004a63b8 FcConfigAddDirList + 157
2 libfontconfig.1.dylib 0x004a6471 FcConfigBuildFonts + 123
3 libfontconfig.1.dylib 0x004b1dc0 FcInitLoadConfigAndFonts + 45
4 libfontconfig.1.dylib 0x004b1e0c FcInit + 41
5 libfontconfig.1.dylib 0x004a650a FcConfigGetCurrent + 30
6 libfontconfig.1.dylib 0x004a8302 FcConfigSubstituteWithPat + 25
7 libfontconfig.1.dylib 0x004a8a07 FcConfigSubstitute + 39
8 libtk8.5.dylib 0x001c3613 InitFont + 72
...

(This is 32bit mode. In 64bit mode it made it into freetype before crashing.)

Unfortunately, I think tcltk/X11 on OSX is forced to used the X11 freetype/fontconfig build, instead of the one I configure for GRASS (my freetype framework).

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

Theory of the Universe

There is a theory which states that if ever anyone discovers exactly what the universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarrely inexplicable. There is another theory which states that this has already happened.

-Hitchhiker's Guide to the Galaxy 2nd season intro

8

Ah, it looks like --enable-xft is a new option in Tk 8.5, and is enabled by default. In the working Tiger TclTk 8.5 GUI, the menu fonts look nice, compared to the 8.4 GUI.

So, disabled xft and now TclTk 8.5 GUI works on Leopard. NVIZ also works. NVIZ still won't run on Tiger though.

On Mar 9, 2008, at 6:01 PM, William Kyngesburye wrote:

Some initial tests for TclTk 8.5.1 on OSX.

10.4/Tiger: the GUI runs and seems to work properly. NVIZ won't run - with Spearfish:

"
nviz elev=elevation.10m

alloc: invalid block: 0x5202b4: 61 0 0

[2]+ Abort trap
"

Even just "nviz --help" gives that alloc error.

10.5/Leopard: won't run at all. There appears to be some problem between Leopard's fontconfig/freetype and TclTk 8.5 (TclTk 8.4 works).

Thread 0 Crashed:
0 libfontconfig.1.dylib 0x004a620a FcConfigAddCache + 149
1 libfontconfig.1.dylib 0x004a63b8 FcConfigAddDirList + 157
2 libfontconfig.1.dylib 0x004a6471 FcConfigBuildFonts + 123
3 libfontconfig.1.dylib 0x004b1dc0 FcInitLoadConfigAndFonts + 45
4 libfontconfig.1.dylib 0x004b1e0c FcInit + 41
5 libfontconfig.1.dylib 0x004a650a FcConfigGetCurrent + 30
6 libfontconfig.1.dylib 0x004a8302 FcConfigSubstituteWithPat + 25
7 libfontconfig.1.dylib 0x004a8a07 FcConfigSubstitute + 39
8 libtk8.5.dylib 0x001c3613 InitFont + 72
...

(This is 32bit mode. In 64bit mode it made it into freetype before crashing.)

Unfortunately, I think tcltk/X11 on OSX is forced to used the X11 freetype/fontconfig build, instead of the one I configure for GRASS (my freetype framework).

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

Theory of the Universe

There is a theory which states that if ever anyone discovers exactly what the universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarrely inexplicable. There is another theory which states that this has already happened.

-Hitchhiker's Guide to the Galaxy 2nd season intro

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"History is an illusion caused by the passage of time, and time is an illusion caused by the passage of history."

- Hitchhiker's Guide to the Galaxy

9

Maybe this would help on windows as well? Currently, you have to install
tcltk 8.4, to get nviz working

jachym

William Kyngesburye píše v Ne 09. 03. 2008 v 19:52 -0500:

Ah, it looks like --enable-xft is a new option in Tk 8.5, and is
enabled by default. In the working Tiger TclTk 8.5 GUI, the menu
fonts look nice, compared to the 8.4 GUI.

So, disabled xft and now TclTk 8.5 GUI works on Leopard. NVIZ also
works. NVIZ still won't run on Tiger though.

On Mar 9, 2008, at 6:01 PM, William Kyngesburye wrote:

> Some initial tests for TclTk 8.5.1 on OSX.
>
> 10.4/Tiger: the GUI runs and seems to work properly. NVIZ won't run
> - with Spearfish:
>
> "
> nviz elev=elevation.10m
>
> alloc: invalid block: 0x5202b4: 61 0 0
>
> [2]+ Abort trap
> "
>
> Even just "nviz --help" gives that alloc error.
>
>
> 10.5/Leopard: won't run at all. There appears to be some problem
> between Leopard's fontconfig/freetype and TclTk 8.5 (TclTk 8.4 works).
>
> Thread 0 Crashed:
> 0 libfontconfig.1.dylib 0x004a620a FcConfigAddCache + 149
> 1 libfontconfig.1.dylib 0x004a63b8 FcConfigAddDirList + 157
> 2 libfontconfig.1.dylib 0x004a6471 FcConfigBuildFonts + 123
> 3 libfontconfig.1.dylib 0x004b1dc0
> FcInitLoadConfigAndFonts + 45
> 4 libfontconfig.1.dylib 0x004b1e0c FcInit + 41
> 5 libfontconfig.1.dylib 0x004a650a FcConfigGetCurrent + 30
> 6 libfontconfig.1.dylib 0x004a8302
> FcConfigSubstituteWithPat + 25
> 7 libfontconfig.1.dylib 0x004a8a07 FcConfigSubstitute + 39
> 8 libtk8.5.dylib 0x001c3613 InitFont + 72
> ...
>
> (This is 32bit mode. In 64bit mode it made it into freetype before
> crashing.)
>
> Unfortunately, I think tcltk/X11 on OSX is forced to used the X11
> freetype/fontconfig build, instead of the one I configure for GRASS
> (my freetype framework).
>
> -----
> William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
> http://www.kyngchaos.com/
>
> Theory of the Universe
>
> There is a theory which states that if ever anyone discovers exactly
> what the universe is for and why it is here, it will instantly
> disappear and be replaced by something even more bizarrely
> inexplicable. There is another theory which states that this has
> already happened.
>
> -Hitchhiker's Guide to the Galaxy 2nd season intro
>
>
> _______________________________________________
> grass-dev mailing list
> grass-dev@lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/grass-dev

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"History is an illusion caused by the passage of time, and time is an
illusion caused by the passage of history."

- Hitchhiker's Guide to the Galaxy

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

--
Jachym Cepicky
e-mail: jachym.cepicky gmail com
URL: http://les-ejk.cz
GPG: http://www.les-ejk.cz/pgp/jachym_cepicky-gpg.pub

10

On Mon, 10 Mar 2008, Jachym Cepicky wrote:

Maybe this would help on windows as well? Currently, you have to install
tcltk 8.4, to get nviz working

Really? NVIZ with Tcl/Tk 8.5.0 on Windows works fine for me, as I've already reported to the list several times. Benjamin reported that it hung for him when adding a vector points layer, but I have no problem with that. Is there maybe a problem with 8.5.1?

However, in a bit of further testing this morning I've discovered a new error that appears from different sources:

couldn't duplicate input handle: bad file number
   while executing
"exec $src_boot/bin/g.gisenv GISBASE"
   (procedure "Nv_makeGUI" line 13)
   invoked from within
"Nv_makeGUI.top"
   (file "s:/grass/trunk/dist.i686-pc-mingw32\etc\nviz2.2\scripts\nviz2.2_script" line 1025)

But the same error occurs in other places too - notably the call to g.gisenv in grass_file_name{} in mapBrowser.tcl. So I think it is some general Windows issue (possibly also related to Tcl 8.5). Maybe it's a Tcl bug or maybe the calls to exec need to be organised in a different way.

No more time to look into it right now but the next thing I'd like to do would be to check the Tcl release notes/changelog for any changes related to Win32 pipes handling, and maybe try the same thing twith 8.5.1

Paul

11

OK, I have checked a little bit for documentation on how to use
the new, more "native-looking" widgets on TclTk8.5.

Documentation is sparse, to say the least.
It appears that using the new widgets would essentially mean rewriting
large parts of the GUI code:

http://wiki.tcl.tk/15443

Unfortunately, there is no such thing as a drop-in replacement for
the old, ugly widgets.

Thus, updating the TclTk support to make use of the cool new 8.5
look & feel would force us to create an 8.5 branch and maintain it.

Of course, other new things, such as increased speed would come
for free ...

As far as NVIZ is concerned: it freezes and crashes all over the
place if started from the MSYS shell. I have yet to narrow down the
problems a bit more.

Benjamin

Paul Kelly wrote:

On Mon, 10 Mar 2008, Jachym Cepicky wrote:

Maybe this would help on windows as well? Currently, you have to install
tcltk 8.4, to get nviz working

Really? NVIZ with Tcl/Tk 8.5.0 on Windows works fine for me, as I've already reported to the list several times. Benjamin reported that it hung for him when adding a vector points layer, but I have no problem with that. Is there maybe a problem with 8.5.1?

However, in a bit of further testing this morning I've discovered a new error that appears from different sources:

couldn't duplicate input handle: bad file number
  while executing
"exec $src_boot/bin/g.gisenv GISBASE"
  (procedure "Nv_makeGUI" line 13)
  invoked from within
"Nv_makeGUI.top"
  (file "s:/grass/trunk/dist.i686-pc-mingw32\etc\nviz2.2\scripts\nviz2.2_script" line 1025)

But the same error occurs in other places too - notably the call to g.gisenv in grass_file_name{} in mapBrowser.tcl. So I think it is some general Windows issue (possibly also related to Tcl 8.5). Maybe it's a Tcl bug or maybe the calls to exec need to be organised in a different way.

No more time to look into it right now but the next thing I'd like to do would be to check the Tcl release notes/changelog for any changes related to Win32 pipes handling, and maybe try the same thing twith 8.5.1

Paul
_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

--
Benjamin Ducke
Senior Applications Support and Development Officer

Oxford Archaeological Unit Limited
Janus House
Osney Mead
OX2 0ES
Oxford, U.K.

Tel.: ++44 (0)1865 263 800
benjamin.ducke@oxfordarch.co.uk

12

I could try it.

On Mar 9, 2008, at 6:36 PM, William Kyngesburye wrote:

I can build an OSX GRASS app with TclTk 8.5 for others to test, and try it myself.

If we intend to support multiple versions of TclTk, it would be helpful to have a TCLTKVER variable in Platform.make for the configured major.minor version.

On Mar 8, 2008, at 9:06 AM, Michael Barton wrote:

IMO we should at least attempt to support Tcl/Tk 8.3 - 8.5.

For 8.3 there is just the one problem with v.digit, AFAIK.

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

First Pogril: Why is life like sticking your head in a bucket filled with hyena offal?
Second Pogril: I don't know. Why IS life like sticking your head in a bucket filled with hyena offal?
First Pogril: I don't know either. Wretched, isn't it?

-HitchHiker's Guide to the Galaxy

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

13

That would be great. Probably not worth worrying about the new
widgets right now.
Rather important to get NVIZ and the remaining MacOS X issues
sorted out for 8.5.1.

Agustin Diez Castillo wrote:

I could try it.

On Mar 9, 2008, at 6:36 PM, William Kyngesburye wrote:

I can build an OSX GRASS app with TclTk 8.5 for others to test, and try it myself.

If we intend to support multiple versions of TclTk, it would be helpful to have a TCLTKVER variable in Platform.make for the configured major.minor version.

On Mar 8, 2008, at 9:06 AM, Michael Barton wrote:

IMO we should at least attempt to support Tcl/Tk 8.3 - 8.5.

For 8.3 there is just the one problem with v.digit, AFAIK.

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

First Pogril: Why is life like sticking your head in a bucket filled with hyena offal?
Second Pogril: I don't know. Why IS life like sticking your head in a bucket filled with hyena offal?
First Pogril: I don't know either. Wretched, isn't it?

-HitchHiker's Guide to the Galaxy

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

--
Benjamin Ducke
Senior Applications Support and Development Officer

Oxford Archaeological Unit Limited
Janus House
Osney Mead
OX2 0ES
Oxford, U.K.

Tel.: ++44 (0)1865 263 800
benjamin.ducke@oxfordarch.co.uk

14

On Mon, 10 Mar 2008, Benjamin Ducke wrote:

As far as NVIZ is concerned: it freezes and crashes all over the
place if started from the MSYS shell. I have yet to narrow down the
problems a bit more.

Hmm, I have only tried starting it from the Windows cmd.exe shell. Perhaps starting from the Msys shell adds a level of complexity. The few errors I *did* get seemed related to starting and running sub-processes, so there could be a connection there.

Paul

15

Probably problem is mixed back and forward slashes within single call?
Please, open new ticket for this (hardcoded paths in nviz).

Maris.

2008/3/10, Paul Kelly <paul-grass@stjohnspoint.co.uk>:

However, in a bit of further testing this morning I've discovered a new
error that appears from different sources:

couldn't duplicate input handle: bad file number
while executing
"exec $src_boot/bin/g.gisenv GISBASE"
(procedure "Nv_makeGUI" line 13)
invoked from within
"Nv_makeGUI.top"
(file
"s:/grass/trunk/dist.i686-pc-mingw32\etc\nviz2.2\scripts\nviz2.2_script"
line 1025)

Paul

16

Hin

NVIZ works fine for me, with tcl/tk 8.5.1
BTW, I just finished the installer for windows, but there’s still a problem I’m not able to fix:

installer procedure creates 4 different files to let GRASS work even if installed in dirs with spaces; files are:
grass63.bat
grass63 (to launch grass from msys shell)
msys.bat
grassrc6 (to avoid initial location error)

It works fine, but nviz fails if GRASS is installed in a “spaced” dir; I verified that error doesn’t occur when installed in a “space free” dir!

I could release it right now (actually, after lunch ;-)), warning in the install procedure to not use dir with spaces, or delay the release and try to fix the problem.
What do you think about?

Marco

Da: grass-dev-bounces@lists.osgeo.org per conto di Maris Nartiss
Inviato: lun 10/03/2008 13.14
A: GRASS developers list
Oggetto: Re: [GRASS-dev] Re: GUI support for TclTk 8.5

Probably problem is mixed back and forward slashes within single call?
Please, open new ticket for this (hardcoded paths in nviz).

Maris.

2008/3/10, Paul Kelly paul-grass@stjohnspoint.co.uk:

However, in a bit of further testing this morning I’ve discovered a new
error that appears from different sources:

couldn’t duplicate input handle: bad file number
while executing
“exec $src_boot/bin/g.gisenv GISBASE”
(procedure “Nv_makeGUI” line 13)
invoked from within
“Nv_makeGUI.top”
(file
“s:/grass/trunk/dist.i686-pc-mingw32\etc\nviz2.2\scripts\nviz2.2_script”
line 1025)

Paul

grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

17

On Mon, 10 Mar 2008, Maris Nartiss wrote:

Probably problem is mixed back and forward slashes within single call?

No, there are no mixed slashes - the paths output by g.gisenv for GISBASE and GISDBASE all contain forward slashes. I think Windows Tcl/Tk can handle mixed forward and back fine anyway. Windows system calls certainly can.

Please, open new ticket for this (hardcoded paths in nviz).

I'm not sure enough what the problem is to do that. If I knew what it was, I'd try to fix it! As I said, I *suspect* it's not directly related to this code (which has been unchanged for years).

Paul

18

On Mon, 10 Mar 2008 marco.pasetti@alice.it wrote:

Hin

NVIZ works fine for me, with tcl/tk 8.5.1
BTW, I just finished the installer for windows, but there's still a problem I'm not able to fix:

installer procedure creates 4 different files to let GRASS work even if installed in dirs with spaces; files are:
grass63.bat
grass63 (to launch grass from msys shell)
msys.bat
grassrc6 (to avoid initial location error)

It works fine, but nviz fails if GRASS is installed in a "spaced" dir; I verified that error doesn't occur when installed in a "space free" dir!

I could release it right now (actually, after lunch ;-)), warning in the install procedure to not use dir with spaces, or delay the release and try to fix the problem.
What do you think about?

How does NVIZ fail? Is there an error message? We could try to fix it.

Installing in a directory with spaces is never a good idea. There are various little things here and there that will break, although the situation is better than it used to be. In particular, I think shell scripts invoked with the batch file wrapper to GRASS_SH (normally set to the Msys shell) will fail because of a documented bug in the Windows _spawnlp() function which is used in g.parser. I used to have a fix for this (I think it just involved some extra quoting of the script name) in my local source tree but I never committed it because I wasn't very sure if it was the best solution, and I think I have lost it now in the move from CVS to SVN.

Paul

19

I put an OSX SVN build on my GRASS page with TclTk 8.5. NVIZ works on Leopard but not on Tiger. It installs in /Applications/GRASS-SVN so it doesn't overwrite the RC5 release.

http://www.kyngchaos.com/wiki/software:grass

For further testing, it also includes the wxpython vdigit that seems to be working.

On Mar 10, 2008, at 6:07 AM, Agustin Diez Castillo wrote:

I could try it.

On Mar 9, 2008, at 6:36 PM, William Kyngesburye wrote:

I can build an OSX GRASS app with TclTk 8.5 for others to test, and try it myself.

If we intend to support multiple versions of TclTk, it would be helpful to have a TCLTKVER variable in Platform.make for the configured major.minor version.

On Mar 8, 2008, at 9:06 AM, Michael Barton wrote:

IMO we should at least attempt to support Tcl/Tk 8.3 - 8.5.

For 8.3 there is just the one problem with v.digit, AFAIK.

For 8.5 it apparently mostly works, if it is just a few little issues
which are not so hard to fix, then we should try and solve them.
i.e. we should at least look into 8.5 problems, it is probably just 1-5
lines of code to be adjusted, and that's not a huge drain on
development resources.

You are right that fixing issues is probably pretty simple. It's figuring out what to fix that is time consuming. Right now, I'm heavily overcommitted timewise. And on top of that, I'm using my laptop (primary development box) for my GIS class and can't afford to have GRASS broken for more than a couple days. So I have no problem in theory, fixing the GUI for 8.5 if it doesn't involve NVIZ internals or something like that, but won't have time to do much for awhile.

Michael

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

First Pogril: Why is life like sticking your head in a bucket filled with hyena offal?
Second Pogril: I don't know. Why IS life like sticking your head in a bucket filled with hyena offal?
First Pogril: I don't know either. Wretched, isn't it?

-HitchHiker's Guide to the Galaxy

_______________________________________________
grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

Theory of the Universe

There is a theory which states that if ever anyone discovers exactly what the universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarrely inexplicable. There is another theory which states that this has already happened.

-Hitchhiker's Guide to the Galaxy 2nd season intro

20

Tiger or Leopard?

On Mar 10, 2008, at 12:31 PM, Agustin Diez Castillo wrote:

It's working here, nice and easy<Picture 10.png>

On Mar 10, 2008, at 4:19 PM, William Kyngesburye wrote:

I put an OSX SVN build on my GRASS page with TclTk 8.5. NVIZ works on Leopard but not on Tiger. It installs in /Applications/GRASS-SVN so it doesn't overwrite the RC5 release.

http://www.kyngchaos.com/wiki/software:grass

For further testing, it also includes the wxpython vdigit that seems to be working.

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

"Time is an illusion - lunchtime doubly so."

- Ford Prefect