[GRASS-stats] [R-sig-Geo] GISRC environment variable using initGRASS

Mon Apr 18 04:28:48 PDT 2016

On Mon, 18 Apr 2016, Roger Bär wrote:

> ----------------------------------------
>> Date: Mon, 18 Apr 2016 10:40:02 +0200
>> From: Roger.Bivand at nhh.no
>> To: baerroger at outlook.com
>> CC: grass-stats at lists.osgeo.org
>> Subject: RE: [R-sig-Geo] GISRC environment variable using initGRASS
>>
>> On Mon, 18 Apr 2016, Roger Bär wrote:
>>
>>>
>>>
>>> ----------------------------------------
>>>> Date: Sun, 17 Apr 2016 20:11:06 +0200
>>>> From: Roger.Bivand at nhh.no
>>>> To: baerroger at outlook.com
>>>> CC: grass-stats at lists.osgeo.org
>>>> Subject: RE: [R-sig-Geo] GISRC environment variable using initGRASS
>>>>
>>>> On Sat, 16 Apr 2016, Roger Bär wrote:
>>>>
>>>>> Thanks for the support!Please see below my comments.
>>>>>
>>>>>> Date: Fri, 15 Apr 2016 20:58:15 +0200
>>>>>> From: Roger.Bivand at nhh.no
>>>>>> To: baerroger at outlook.com
>>>>>> CC: r-sig-geo at r-project.org
>>>>>> Subject: Re: [R-sig-Geo] GISRC environment variable using initGRASS
>>>>>>
>>>>>> On Fri, 15 Apr 2016, Roger Bär wrote:
>>>>>>
>>>>>>> I am currently using the R package "rgrass7" (Version: 0.1-6) in order
>>>>>>> to command GRASS GIS (Windows7, OSGEO4W installation).
>>>>>>>
>>>>>>> I use R in order to start a GRASS environment and to script my
>>>>>>> geoprocessing. Both is working fine. However, I have problems with
>>>>>>> storing and calling the GRASS gisenv variables. Every time when I call
>>>>>>> "initGRASS" a file called "junk" is created containing GRASS gisenv
>>>>>>> variables.
>>>>>>>
>>>>>>
>>>>>> What the file is called is probably immaterial.
>>>>>>
>>>>>>>
>>>>>>> Looking at the source of the initGRASS function the section below
>>>>>>> puzzeled me somehow:
>>>>>>>
>>>>>>> 59 Sys.setenv(GISRC = paste(Sys.getenv("HOME"), "\\.grassrc7",
>>>>>>> 60 sep = ""))
>>>>>>
>>>>>> This is setting GISRC to a value
>>>>>
>>>>> Yes. The absolute path to the file (containing settings for GISDBASE,
>>>>> LOCATION_NAME and MAPSET)
>>>>>
>>>>>>> 61 if (file.exists(Sys.getenv("GISRC")) && !override)
>>>>>>> 62 stop("A GISRC file already exists; to override, set override=TRUE")
>>>>>>> 63 Sys.setenv(GISRC = "junk")
>>>>>>
>>>>>> This is setting it again in the working directory - I don't recall why -
>>>>>> probably frustration years ago with Windows.
>>>>>
>>>>> I really do not understand the "junk" (line 63). Because GISRC is set to
>>>>> "junk" the subsequent lines of code will write the GISRC environment
>>>>> variables to a newly created file called "junk" in the home folder.
>>>>> Shouldn't this line not rather be omitted? (The GRASS environment
>>>>> variables would then be written to ".grassrc7", which in my opinion
>>>>> would be correct)
>>>>
>>>> The name of the file is immaterial, but from an R script should only be
>>>> accessed from the value of the environment variable.
>>>>
>>>
>>> Ok.
>>>
>>>>>
>>>>>>
>>>>>>> 64 cat("GISDBASE:", getwd(), "\n", file = Sys.getenv("GISRC"))
>>>>>>> 65 cat("LOCATION_NAME: <UNKNOWN>", "\n", file = Sys.getenv("GISRC"),
>>>>>>> 66 append = TRUE)
>>>>>>> 67 cat("MAPSET: <UNKNOWN>", "\n", file = Sys.getenv("GISRC"),
>>>>>>> 68 append = TRUE)
>>>>>>> 69 gisrc <- ifelse(use_g.dirseps.exe, system(paste("g.dirseps.exe -g",
>>>>>>> 70 shQuote(Sys.getenv("GISRC"))), intern = TRUE), Sys.getenv("GISRC"))
>>>>>>> 71 assign("addEXE", .addexe(), envir = .GRASS_CACHE)
>>>>>>> 72 Sys.setenv(GISRC = gisrc)
>>>>>>>
>>>>>>
>>>>>> And this sets it again. It is set, otherwise GRASS wouldn't work;
>>>>>> programmatically you'd have to use Sys.getenv("GISRC") to get the value.
>>>>>> Please provide a use case where this matters, and consider moving this
>>>>>> thread to the statsgrass list.
>>>>>
>>>>> Sorry, I do not understand what you mean with "provide a use case where
>>>>> this matters". What do you mean with "this"?
>>>>
>>>> The use of an arbitrary name in a throw-away location is of no
>>>> significance, and could in the future be a short random name. My guess is
>>>> that the use of "junk" may have come from different file permissions on
>>>> unix and windows systems - if override is TRUE, R might end up trying to
>>>> write to a file that some other process is still blocking.
>>>>
>>>
>>> Ok.
>>>
>>>> Unless you can provide an R script that I can run in WinNat (not OSGeo4W -
>>>> I do not have the time or patience to install multiple Windows versions of
>>>> GRASS that I never use), and which demonstrates that something more is
>>>> involved than the name of a temporary file, I can't help, because I do not
>>>> see any real problem - the name if the file is arbitrary, and in the R
>>>> session is accesses by Sys.getenv().
>>>
>>> Ok, I get that. The main issue was to keept the directory clean since
>>> access my grass data from different operating system. But thinks it's
>>> not worth to spend more time on this issue. (I'll survive the junk file
>>> in my workspace)
>>
>> Would it help to try to add an argument to initGRASS() to try to delete
>> the file pointed at by GISRC on exit from the R session?
>>
>> In many cases, the file is in R's temporary directory (home=tempdir()), so
>> gets removed automatically, but if you are using a non-temporary home=,
>> the file stays there.
>>
>> Roger (Bi)
>
> Yes, that would solve my issue indeed.  Because I am using initGRASS 
> mostly in a permanent home with all my project data.

Please download a development version from:

http://win-builder.r-project.org/JZw14G61RjVk

Set remove_GISRC=TRUE in the initGRASS() call, and use remove_GISRC() when 
you are through with this instance. I have included code to try to unlink 
the file and unset the environment variable when the R session terminates, 
but so far without success - you will need to remove it as described 
unless I see a way round a race condition.

Roger (Bi)

>
> Roger (Bä)
>
>
>>> Thanks you anyway!
>>>
>>>
>>> Regards,
>>> Roger (Bä)
>>>
>>>
>>>
>>>
>>>>> I added the thread to the grass-stat list (and will remove r-sig from
>>>>> the cc the next time).
>>>>
>>>> OK,
>>>>
>>>> Roger (Bi)
>>>>
>>>>> Roger
>>>>>
>>>>>>
>>>>>> Hope this helps (doesn't clarify, but that's legacy Windows),
>>>>>>
>>>>>> Roger
>>>>>>
>>>>>>> First, why is initGRASS using ".grassrc7" for storing the GISRC file?
>>>>>>> InitGRASS is looking for the GISRC file under ".grassrc7" (line 59). However, in the package documentation [1], the GISRC file is called ".gisrc". Moreover, the GRASS documentation says that the GISRC file is stored under "/.grass7/rc"
>>>>>>>
>>>>>>> Second, what is the purpose of assigning "junk" to the GISRC environment variable?
>>>>>>> GISRC is set to "junk" (line 63) and then the GISRC value ("junk") is assigned again back to system variable GISRC. Unless I misinterpreded the code, it does not make much sense to me. Shouldn't the enviroment varibales not be written (line 64 - 68) to ".grassrc7" rather than to "junk"?
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> I would appreciate any support or hints!
>>>>>>>
>>>>>>> Regards,
>>>>>>> Roger
>>>>>>>
>>>>>>> [1] http://https//cran.r-project.org/web/packages/rgrass7/rgrass7.pdf
>>>>>>> [2] http://grass.osgeo.org/grass70/manuals/variables.html
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> R-sig-Geo mailing list
>>>>>>> R-sig-Geo at r-project.org
>>>>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>>>>>>
>>>>>>
>>>>>> --
>>>>>> Roger Bivand
>>>>>> Department of Economics, Norwegian School of Economics,
>>>>>> Helleveien 30, N-5045 Bergen, Norway.
>>>>>> voice: +47 55 95 93 55; fax +47 55 95 91 00
>>>>>> e-mail: Roger.Bivand at nhh.no
>>>>>> http://orcid.org/0000-0003-2392-6140
>>>>>> https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en
>>>>>> http://depsy.org/person/434412
>>>>>
>>>>
>>>> --
>>>> Roger Bivand
>>>> Department of Economics, Norwegian School of Economics,
>>>> Helleveien 30, N-5045 Bergen, Norway.
>>>> voice: +47 55 95 93 55; fax +47 55 95 91 00
>>>> e-mail: Roger.Bivand at nhh.no
>>>> http://orcid.org/0000-0003-2392-6140
>>>> https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en
>>>> http://depsy.org/person/434412
>>
>> --
>> Roger Bivand
>> Department of Economics, Norwegian School of Economics,
>> Helleveien 30, N-5045 Bergen, Norway.
>> voice: +47 55 95 93 55; fax +47 55 95 91 00
>> e-mail: Roger.Bivand at nhh.no
>> http://orcid.org/0000-0003-2392-6140
>> https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en
>> http://depsy.org/person/434412

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 91 00
e-mail: Roger.Bivand at nhh.no
http://orcid.org/0000-0003-2392-6140
https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en
http://depsy.org/person/434412