svSocket and R-devel: possible bug
jcfaria opened this issue · comments
Hi,
I am writing to inform you that a possible bug has been detected in the svSocket (1.0.2) package in the R-devel version (4.2.0). I tested it on three machines: 1) my local Windows 10 machine; 2) virtual machine with Windows Server 2019 on AWS; 3) virtual machine with Windows 10 professional on Azure platform.
The error message is:
Error in assign("last.warning", last.warning, envir = baseenv()) :
cannot add binding of 'last.warning' to the base environment
About two years ago a similar error message was detected in Microsoft R Open and the problem was solved with the instruction that can be seen (see attached figure:
) in line 33 of the info.R file which is open in Tinn-R . However, it is not working for R-devel.
Workaround bug: MRO and svSocket
unlockBinding("last.warning",
baseenv())
Is it possible to know if the bug is in R or in the svSocket package? How to fix?
Dear Phil,
Here I tested version 1.2.1 of the svMisc package (as dependent on the main svSocket package) on R-devel (4,2,0) and MRO (4.0.2): I can say that it ran well, I didn't see any problems. Even the new svSocket version allowed removing from the info.R file of Tinn-R project the way the problem was worked around in the MRO.
#. Workaround bug: MRO and svSocket
#.unlockBinding("last.warning",
#. baseenv())
I am grateful for the attention as always.
All the best.
Excellent! I plan more fixes and additions to {svMisc} until the end of July. So, I would not submit it to CRAN before that. In the meantime, release 1.2.1 of {svMisc} can be installed with remotes::install_github("SciViews/svMisc@v1.2.1")
. It is identical to the current CRAN version except for this fix and two new functions to access a pCloud drive, that I think, you don't use.