Marxan input files (pu.dat, etc.) are being written to project main dir

[langfob]

Not sure why, but pu.dat, puvspr.dat, and spec.dat are showing up in the main directory for the package (alongside the R directory, NAMESPACE file, etc.).

Has it always been doing this or is this a new bug?

Regardless, they should be going somewhere else.

Note that this is also listed as an issue in bdpgxupaper, so close it there too when this one closes.

In reading Rappdirs section of "Mastering software development in R", it mentions the rappdirs package and after loading that and entering ??rappdirs, I came across a quote from the CRAN policies that says:

Packages should not write in the users' home filespace, nor anywhere else on the file system apart from the R session's temporary directory (or during installation in the location pointed to by TMPDIR: and such usage should be cleaned up). Installing into the system's R installation (e.g., scripts to its bin directory) is not allowed.

Need to looks some more at this rappdirs package to see if it helps...

[langfob]

These files are currently ending up in the bdpgxupaper R directory instead of in some tzar output area. Are these just from runs that crashed during development? They seem to be from a couple of months ago.

[langfob]

Going through my daily log file, I also found an old entry about where these files go. I'll just paste that entry in here in case it still applies:

2015-03-08 Sunday

• Tried to make all of the marxan input and output files write to the tzar output area instead of to the directory where the marxan executable is. Succeeded with all except the input.dat file.
• Could not figure out how to get input.dat to be read from the tzar area rather than marxan’s home directory.
• I can’t find any documentation on the web or in the marxan manuals about what command line arguments are allowed.
• I went through every letter of the alphabet as a command line option and only –s and –c were not rejected by marxan.
  • I have no idea what –c does. Can’t find any documentation about it.
  • Also, no variations on –h or –help or --h or --help worked, nor does anything like –v or –version.
• So, unless I can figure it out, I'm going to have to write the input.dat file to wherever the marxan executable is and then copy it over to the marxan_io area after it's written out.

[langfob]

In an email or a document sometime in the last few months, Ascelin also mentioned something about marxan input files ending up in the wrong place in some of his test runs. Need to find that again to see if it relates to all this too.

[langfob]

Found another daily log entry that may be relevant:

2015-03-26

• Marxan mailing list recently pointed to some new marxan materials on github and one of them has an example of some R code that calls marxan with what looks like a specification of the location of the input.dat file.
  • So, it looks like I can just add the filespec of the input.dat file as an argument after the "-s" argument without having any kind of other dash option specifying that you're giving the path to the input.dat file.
  • See around line 406 of:
    • mattwatts/MarxanRandomStuff@281b605
  • There, he shows a set of calls to marxan using files called input1.dat, input2.dat, etc.