Skip to content

[R-pkg-devel] How to fix Archived Package Rpdb?

6 messages · Hadley Wickham, Uwe Ligges, Leonard Mada +1 more

Original
Leonard Mada
# 
Dear Members,

I would like to reanimate the archived package Rpdb:
https://cran.r-project.org/web/packages/Rpdb/index.html

1.) I have tried to contact the original author by email, but got no 
response.

2.) New Repository on GitHub
I have copied the existing code to a new repository on GitHub:
https://github.com/discoleo/Rpdb

- fixed the use of deprecated functions (rgl);
- fixed some bug with Roxygen2;
- I hope that all errors are now fixed;

2.b.) Description file
- I left the original author as the author (with the provided e-mail 
address): should I delete this email?
- I have added myself as maintainer;
- I have increased the last digit of the version number;
- I have added links to this new GitHub repository: I did not find any 
other links in the previous version (except to the pdb-format);
- updated the licence to GPL v3: the original does not specify any 
version number;


Is there anything else that needs to be done?


I may wait another 1 or 2 days and will try to submit the new version of 
the package.


Many thanks,


Leonard
Hadley Wickham
# 
On Fri, Sep 8, 2023 at 6:02?AM Leonard Mada via R-package-devel
<r-package-devel at r-project.org> wrote:
It probably doesn't matter than much either way, but since the author
doesn't appear to respond to emails to that address, I personally
would lean towards deleting it.
There are at least three 3 R CMD check failures you need to address:

* The Authors at R field in DESCRIPTION is incorrectly formed, you need
something like this:
    c(
      person("Leonard", "Mada", email = "leo.mada at syonic.eu", role = c("cre")),
      person("Julien", "Id?", role = c("aut"))
    )

* You need to add LICENSE to .Rbuildignore, or and IMO better, delete
that file and use usethis::use_gpl3_license() to the license in
markdown form, and correctly ignored for CRAN submission

* Many examples use `\%in\%` instead of `%in%.

To make these sorts of problems easier to spot in the future I'd
suggest setting up a GitHub action to automatically run R CMD check
every time you push to GitHub. One easy way to do that is to run
usethis::use_github_action("check-standard").

Hadley
1 day later
Leonard Mada
# 
Thank you very much for this help.


1.) I am a little bit unsure about the LICENSE file - see below (in-text).


2.) There is a new error in the meantime:

- the check works on Windows, but fails everywhere else with:

Warning: Found the following significant warnings:
 ? Warning: 'rgl.init' failed, running with 'rgl.useNULL = TRUE'.


Googling the web was not very informative either: it mentions something 
about quartz device - but I am uncertain what to do.
On 9/8/2023 6:59 PM, Hadley Wickham wrote:
If I understand correctly:

- delete the "LICENSE" file and use usethis::use_gpl3_license(), which 
adds the "LICENSE.md" file;

- should I also add some code to the DESCRIPTION file?
Hopefully, this is fixed now. But it was quit a hassle to find out which 
files were affected. [I could have used gawk, but the error-reporting 
could be improved as well!]


Sincerely,


Leonard
2 days later
Uwe Ligges
#
On 09.09.2023 20:15, Leonard Mada via R-package-devel wrote:
On non Windows systems, You cannot use rgl if you do not have any X11 
available.
Support for Unix alikes is optional, so in packages X11() should be used 
conditionally after checking capabilities("X11").
>
Do *not* delete any authours/copyright holders.
LICENSE: GPL-3

in the DESCRIPTION should be fine, and no license file unless you want 
to add additional restrictions that are permitted by GPL-3 such as 
attribution requirements.

No idea what usethis::use_gpl3_license() does.

Best,
Uwe Ligges
1 day later
Leonard Mada
# 
Thank you very much for your response.

I actually tried yesterday to publish the package to CRAN. 
Unfortunately, my computer seems to behave funny; and I did not receive 
a confirmatory e-mail either.

Messages & Errors:
? Package submission successful
? Check your email for confirmation link.
! Don't forget to tag this release once accepted by CRAN
Error in (function (..., row.names = NULL, check.rows = FALSE, 
check.names = TRUE,? :
 ? arguments imply differing number of rows: 1, 0
In addition: Warning messages:
1: In system2("git", c("rev-parse", "--abbrev-ref", "HEAD"), stdout = 
TRUE) :
 ? running command '"git" rev-parse --abbrev-ref HEAD' had status 127
2: In system2("git", c("status", "--porcelain=v1"), stdout = TRUE) :
 ? running command '"git" status --porcelain=v1' had status 127
3: In system2("git", c("rev-parse", "HEAD"), stdout = TRUE) :
 ? running command '"git" rev-parse HEAD' had status 127
? Setting active project to 'C:/.../Rpdb'

1.) I do not understand if this means, that the submission actually failed?

2.) Should I try to resubmit?
If yes: should I increase the version number to 2.3.2 (from 2.3.1)?

3.) Old questions
The remaining questions should be sorted out now:
- the issue was if to delete the e-mail from the original author;
- I left the e-mail for now; I will phase out the e-mail address, once I 
start to work on my own on the package (maybe I'll have some time next 
month);

Sincerely,

Leonard
On 9/12/2023 4:02 PM, Uwe Ligges wrote:
Ivan Krylov
# 
13 ???????? 2023 ?. 18:01:56 UTC, Leonard Mada via R-package-devel <r-package-devel at r-project.org> ?????:
This is a bug in devtools where a value originating from subsetting a list using $ (which might be NULL, i.e. of length 0) is compared against a string (i.e. of length 1) using == (which fails) instead of identical() (which would have returned FALSE). Try submitting using the form at https://cran.r-project.org/.
If the package hasn't reached CRAN at all (looks like it hasn't), feel free to keep the version number. Otherwise it's best to increment it.
My personal opinion is that a bouncing e-mail address can be removed. I think that we're all is in agreement that the _name_ of the original author must stay.