[R-pkg-devel] advice about errorCondition() function in the Rcmdr package

-----Original Message-----
From: Uwe Ligges [mailto:ligges at statistik.tu-dortmund.de]
Sent: Wednesday, May 15, 2019 7:51 AM
To: Fox, John <jfox at mcmaster.ca>; R Package Development <r-package-
devel at r-project.org>
Subject: Re: [R-pkg-devel] advice about errorCondition() function in the Rcmdr
package

Dear John,

I guess Rcmdr::errorCondition() will rarely be called by users

(nor
base::errorCondition), hence it should be quite safe if other packages import
correctly into their namespace and use the version their maintainers import?

 From that point of view, there are no changes required in Rcmdr.

Best,
Uwe



On 14.05.2019 16:28, Fox, John wrote:

Dear list members,

I noticed that there's a name clash between errorCondition() in the Rcmdr

package and the new function by the same name in the base package in R
3.6.0. errorCondition() has been in the Rcmdr package for many years and is
used extensively -- literally dozens of times. It's exported because it's also
used in Rcmdr plug-in packages. The function checks for errors, prints an error
message, cleans up after the error, and restarts a dialog.

For the time-being, I've renamed the function as ErrorCondition() (with an

uppercase "E") in the development version of the Rcmdr package, to be
released in late summer, but have retained and deprecated errorCondition()
because I expect that removing it will break most or even all Rcmdr plug-in
packages, but I also suspect that masking base::errorCondition() can be
problematic.

Can anyone suggest the best way to proceed?

Best,
  John

--------------------------------------
John Fox, Professor Emeritus
McMaster University
Hamilton, Ontario, Canada
Web: socialsciences.mcmaster.ca/jfox/