I have taken another step back and tried winGRASS 6.4.4 Command Line.
For Spearfish60 7.2.1 responds with 0xc000007b to `g.proj', `g.list type=rast',
`g.list type=vect', and `v.info soils'. It seems to give a reasonable answer to
`r.info soils'.
6.4.4 responds with 0xc000007b to `g.proj' and `v.info soils' and gives
a reasonable answer to `r.info soils'. It also produces both lists.
Don't the differences let reduce significantly the set of suspect .dlls?
Edmund
On Thu, Aug 17, 2017 at 7:12 PM, Edmund Swylan
<Edmund.Swylan@ponymail.com> wrote:
I have taken another step back and tried winGRASS 6.4.4 Command Line.
For Spearfish60 7.2.1 responds with 0xc000007b to `g.proj', `g.list type=rast',
`g.list type=vect', and `v.info soils'. It seems to give a reasonable answer to
`r.info soils'.6.4.4 responds with 0xc000007b to `g.proj' and `v.info soils' and gives
a reasonable answer to `r.info soils'. It also produces both lists.Don't the differences let reduce significantly the set of suspect .dlls?
Did you try this DLL tool?
https://grasswiki.osgeo.org/wiki/GRASS_Debugging#DLL_dependencies
Markus
ES> I have taken another step back and tried winGRASS 6.4.4 Command Line. ...
ES> Don't the differences let reduce significantly the set of suspect .dlls?
MN> Did you try this DLL tool?
MN> https://grasswiki.osgeo.org/wiki/GRASS_Debugging#DLL_dependencies
I had not tried it. Now I have. To little effect. The dependencies are too
numerous. I have tried to see what makes the v.info of winGRASS 6.4.4
fare worse than the r.info. Taking into consideration what the relevant
.bats do to PATH, I do not see what module might be missed by v.info but
not by r.info (EFSADU.dll, e.g., would be missed by both.). The observation
is not much use though. In the process I seem to have discovered a repetition
in the PATH, a zlib1.dll - probably a copy of zlib_osgeo.dll, and (in 7.2.1)
competing libintl-8.dlls.
Edmund
Edmund Swylan wrote
ES> I have taken another step back and tried winGRASS 6.4.4 Command Line.
...
ES> Don't the differences let reduce significantly the set of suspect
.dlls?
MN> Did you try this DLL tool?
MN> https://grasswiki.osgeo.org/wiki/GRASS_Debugging#DLL_dependenciesI had not tried it. Now I have. To little effect. The dependencies are too
numerous. I have tried to see what makes the v.info of winGRASS 6.4.4
fare worse than the r.info. Taking into consideration what the relevant
.bats do to PATH, I do not see what module might be missed by v.info but
not by r.info (EFSADU.dll, e.g., would be missed by both.). The
observation
is not much use though. In the process I seem to have discovered a
repetition
in the PATH, a zlib1.dll - probably a copy of zlib_osgeo.dll, and (in
7.2.1)
competing libintl-8.dlls
re winGRASS 6.4.4, what I can say, the grass 6.x-line isn't maintained
actively anymore. Thus, testing should be done with the 7.2.x line.
the dll-issues looks like a mismatch and mixing of dlls and their
dependencies in %PATH% variable.
AFAICT starting from scratch may be an option.
-----
best regards
Helmut
--
View this message in context: http://osgeo-org.1560.x6.nabble.com/g-list-1073741701-cntnd-tp5331698p5331889.html
Sent from the Grass - Users mailing list archive at Nabble.com.