-
Notifications
You must be signed in to change notification settings - Fork 9
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
initGrass Error : XML content does not seem to be XML #57
Comments
I cannot reproduce this with GRASS 8.0.1 and 8.0.2 Windows standalones on Windows 10 with R 4.2.0 and rgrass 0.2-10 and XML 3.99-0.9. Could you please run |
The error message:
and the equivalent in the screen shot (please don't post screen shots, copy&paste within a markdown block) shows what the problem is clearly. Neither state whether GRASS is OSGEO4W or Windows native standalone. Both contributors have multiple PROJ installations, where the one being looked for by GRASS is not found. In the second case, the R session has a large number of loaded packages, so it isn't obvious which is the problem (setting
is not protecting |
Please confirm that terra 1.5-21 is the problem. The current release is 1.5-34, for which no error occurs for me. |
The error message is from PROJ detecting a different version of This means in this case that |
The
In terra 1.5-34 the same version (same Rtools PROJ version), but possibly now protected from version conflict.
|
On freshly installed R 4.1.3, back-installing terra 1.5-21, no error is reproducible. |
Thank you, @rsbivand, for your prompt response. Unfortunately, it is tough to track it down. I did all the following:
The only post with the same problem I find (besides this one) is this: https://gis.stackexchange.com/questions/421932/osgeo4w-ogr2ogr-script-with-filegdb-conversion-working-but-getting-a-weird-erro. At this point, it seems that it is not directly related to the grass package. But I would appreciate advice on which GitHub group shall I post this issue to, as many technical terms mentioned are out of my scope of knowledge. Maybe this could help: I tried to run the SQLite script you posted:
It did not work on my system as running the line Or could it be related to QGIS I have installed? I saw other PROJ-related issues in the same context with QGIS. My QGIS version: |
I had the exact same error message on a Windows11 machine. Then I installed standalone GRASS
Instead, it has now been replaced by a warning with the message:
I checked and rechecked to see whether there is any other loaded package that could be loading another |
On Windows, the |
With:
I see:
GRASS 8.2.1 is bundling a slightly older version of |
Here are are the outputs of
If the issue is with the |
Normally, winGRASS, based upon OSGeo4W, uses up-to-date proj versions. Original report talks about GRASS GIS 8.0.2, that is an outdated winGRASS version. |
How often is terra updating proj versions? |
R uses MXE, fairly actively, and is currently working through the cmake shift for PROJ, GDAL and GEOS. The Rtools maintainer updates usually quite early: https://cran.r-project.org/bin/windows/Rtools/rtools43/rtools.html. Status is at https://svn.r-project.org/R-dev-web/trunk/WindowsBuilds/winutf8/ucrt3/toolchain_libs/mxe/src/, https://svn.r-project.org/R-dev-web/trunk/WindowsBuilds/winutf8/ucrt3/toolchain_libs/mxe/src/proj.mk for PROJ. The key thing for Rtools42 and Rtools43 is that they use UCRT, and are future-proof. |
The proj update cycle in OSGeo4W is more or less real time, though standalone installer is not updated. AFAIK no standalone installers of other software are updated in such a way, they are shipped with library versions of that time of creation. So the question may be too If R and packages should be a bit more conservative in updating dependency libraries, as upstream software synchronizing with downstrean R(packages) would be nice, but hardly to achieve, at least in the windows side of the world. In the upcoming GRASS code sprint in the next days, a target will be an attempt to cmake compiling, also on windows. Any help and input welcome. |
Good luck! The major change for R was going for UCRT, which took a lot of work, but which should help in longer-term maintenance (that was an application-wide coming problem). Resolving that took us to MXE and its recipes, and to cross-compiling with recent GCC build trains. Is information about version updates for external software for Windows and/or macOS binaries published anywhere? R packages typically have a function reporting these versions. |
regarding OSGeo4W (and the standalone installers are actually build on that environment): https://download.osgeo.org/osgeo4w/v2/x86_64/setup.ini e.g.
and within a GRASS session
|
Could |
I couldn't find any information of proj-tools giving the info which so I guess: yes it should be inferred from the PROJ version. e.g. see [PROJ] PROJ 9.2.1RC1 & PROJ-data 1.14RC1 |
@rouault thanks, that will help, as both sides can then negotiate about whose |
@hellik When can we expect this |
@fgashakamba please consider contributing, since this is important to you. I am occupied with other topics for the rest of 2023, so contributions are welcome. If there was an easy resolution, it would have been done; since this hasn't happened, the problem is intractable. |
Hello,
I installed GRASS GIS 8.0.2 and tried to open it via R 4.0.3 package rgrass7, but with the following command I got the error message :
initGRASS("C:/Program Files/GRASS GIS 8.0",
override = TRUE,
gisDbase = "GRASS_TEMP",
mapset = "PERMANENT",
location = "test")
Error : XML content does not seem to be XML: 'ontains DATABASE.LAYOUT.VERSION.MINOR = 0 whereas a number >= 2 is expected. It comes from another PROJ installation.'
Also: Warning message:
In system(cmd0, intern = TRUE) :
running the command 'g.proj.exe --interface-description' returns a status of 5
Do you know what is the issue ?
Thank you very much
The text was updated successfully, but these errors were encountered: