[R-pkg-devel] FW: Writing to users config directory for CRAN package

[Jonathan Godfrey]

Further to Dirk's advice, my BrailleR package creates a folder (dumping ground). Users are asked if they want to use  one of my choosing, or a temporary one. If they choose temporary, they get asked again and again until they cave in to my wishes!


BrailleR also writes files to the current working directory, but these are done because the user has chosen to run a command that has the purpose of creating files. I put a warning in the documentation for such functions.

I'm not suggesting my solution is the gold standard, but it is working well enough to keep the CRAN checkers happy.

Jonathan


-----Original Message-----
From: R-package-devel <r-package-devel-bounces using r-project.org> On Behalf Of Dirk Eddelbuettel
Sent: Sunday, 6 November 2022 9:29 am
To: David Hugh-Jones <davidhughjones using gmail.com>
Cc: R package devel <r-package-devel using r-project.org>
Subject: Re: [R-pkg-devel] Writing to users config directory for CRAN package


On 5 November 2022 at 19:32, David Hugh-Jones wrote:
| I'm considering submitting the package onetime (
| https://apc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fhughjonesd%2Fonetime%2F&data=05%7C01%7Ca.j.godfrey%40massey.ac.nz%7C33d5f70186284052580908dabf6c7826%7C388728e1bbd0437898dcf8682e644300%7C1%7C0%7C638032769864581576%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=EbBh9XCyyKsRTACo7eLxASZW3pm%2BTrXxzKjjnxBa%2Fpo%3D&reserved=0) to CRAN.
| 
| Onetime has functions for showing a message or warning only once (ever 
| per user). It does this by writing to a file in the user's 
| configuration directory, as reported by rappdirs::user_config_dir().

There is a base R function tools::R_user_dir() which I use in a few packages along with packageName() to store config information across sessions. A quick search at GitHub's 'cran' org mirroring CRAN finds 110 hits:

   https://apc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fsearch%3Fq%3Dorg%253Acran%2BR_user_dir%26type%3Dcode&data=05%7C01%7Ca.j.godfrey%40massey.ac.nz%7C33d5f70186284052580908dabf6c7826%7C388728e1bbd0437898dcf8682e644300%7C1%7C0%7C638032769864737810%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Wlqbm0tZX0h25rm0veh%2F15IAwJ5mqP9VNPUovlaPhdY%3D&reserved=0

You could keep a hashmap in that directory, and maybe (before it has been written a first time) alert the user that you cannot write without (initial) permission.  As I recall, the idea behind the (sensible) CRAN Policy is to not litter user directories with random files with the user knowing.

Dirk

--
dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org

______________________________________________
R-package-devel using r-project.org mailing list
https://apc01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-package-devel&data=05%7C01%7Ca.j.godfrey%40massey.ac.nz%7C33d5f70186284052580908dabf6c7826%7C388728e1bbd0437898dcf8682e644300%7C1%7C0%7C638032769864737810%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=kB8nHrtC5yGEm4tnOTZDOAGT%2FmtDViGblNvieFlZq7g%3D&reserved=0