总结一下我们的讨论。
1.据我所知，对于base R提供的单元测试，没有R CMD check的标准替代品
1.通常对于单元测试，我在R/下获取所有内容（在source/下获取dyn.load所有内容），然后在tests/下获取所有内容（实际上，我还使用man/目录中帮助页面的Example部分作为测试用例，并将其结果与以前的包版本进行比较）。
我假设这些是devtools和testthat提供的基本测试功能。如果您希望开发多个包并希望独立于非base-R，我建议使用自定义脚本/包自动化上述过程。
我会重新评论一下http://r-pkgs.had.co.nz/tests.html。

测试一个包而不需要

• 运行R CMD check
• 销毁现有安装
• 弄乱你的工作目录

安装到临时目录tmp.lib中，并使用tools::testInstalledPackage在第二个临时目录tmp.out中测试 * 该 * 安装：

pkg.name <- "Matrix"
pkg.root <- normalizePath(file.path("path", "to", pkg.name), mustWork = TRUE) # an absolute path (?)

tmp.lib <- tempfile()
tmp.out <- tempfile()
dir.create(tmp.lib)
dir.create(tmp.out)
owd <- setwd(tmp.out)

tools::Rcmd(c("build", pkg.root))

* checking for file '/x/y/z/path/to/Matrix/DESCRIPTION' ... OK
.
.
.
* building 'Matrix_1.5-5.tar.gz'

list.files(tmp.out)

[1] "Matrix_1.5-5.tar.gz"

tools::Rcmd(c("INSTALL", "-l", tmp.lib, "--install-tests", "*.tar.gz"))

* installing *source* package 'Matrix' ...
.
.
.
* DONE (Matrix)

list.files(tmp.lib)

[1] "Matrix"

tools::testInstalledPackage(pkg.name, lib.loc = tmp.lib, types = "tests")

Running specific tests for package 'Matrix'
  Running 'Class+Meth.R'
  Running 'Simple.R'
  Running 'abIndex-tsts.R'
  Running 'base-matrix-fun.R'
  Running 'bind.R'
  comparing 'bind.Rout' to 'bind.Rout.save' ... OK
  Running 'dg_Matrix.R'
  Running 'dpo-test.R'
  Running 'dtpMatrix.R'
  Running 'factorizing.R'
  Running 'group-methods.R'
  Running 'indexing.R'
  comparing 'indexing.Rout' to 'indexing.Rout.save' ... OK
  Running 'matprod.R'
  Running 'matr-exp.R'
  Running 'other-pkgs.R'
  Running 'packed-unpacked.R'
  Running 'spModel.matrix.R'
  Running 'symmDN.R'
  Running 'validObj.R'
  Running 'write-read.R'

list.files(tmp.out, recursive = TRUE)

[1] "Matrix-tests/Class+Meth.R"         "Matrix-tests/Class+Meth.Rout"     
 [3] "Matrix-tests/Rplots.pdf"           "Matrix-tests/Simple.R"            
 [5] "Matrix-tests/Simple.Rout"          "Matrix-tests/abIndex-tsts.R"      
 [7] "Matrix-tests/abIndex-tsts.Rout"    "Matrix-tests/base-matrix-fun.R"   
 [9] "Matrix-tests/base-matrix-fun.Rout" "Matrix-tests/bind.R"              
[11] "Matrix-tests/bind.Rout"            "Matrix-tests/bind.Rout.save"      
[13] "Matrix-tests/dg_Matrix.R"          "Matrix-tests/dg_Matrix.Rout"      
[15] "Matrix-tests/dpo-test.R"           "Matrix-tests/dpo-test.Rout"       
[17] "Matrix-tests/dtpMatrix.R"          "Matrix-tests/dtpMatrix.Rout"      
[19] "Matrix-tests/factorizing.R"        "Matrix-tests/factorizing.Rout"    
[21] "Matrix-tests/group-methods.R"      "Matrix-tests/group-methods.Rout"  
[23] "Matrix-tests/indexing.R"           "Matrix-tests/indexing.Rout"       
[25] "Matrix-tests/indexing.Rout.save"   "Matrix-tests/matprod.R"           
[27] "Matrix-tests/matprod.Rout"         "Matrix-tests/matr-exp.R"          
[29] "Matrix-tests/matr-exp.Rout"        "Matrix-tests/mm-Matrix.tab"       
[31] "Matrix-tests/other-pkgs.R"         "Matrix-tests/other-pkgs.Rout"     
[33] "Matrix-tests/packed-unpacked.R"    "Matrix-tests/packed-unpacked.Rout"
[35] "Matrix-tests/spModel.matrix.R"     "Matrix-tests/spModel.matrix.Rout" 
[37] "Matrix-tests/symmDN.R"             "Matrix-tests/symmDN.Rout"         
[39] "Matrix-tests/validObj.R"           "Matrix-tests/validObj.Rout"       
[41] "Matrix-tests/write-read.R"         "Matrix-tests/write-read.Rout"     
[43] "Matrix_1.5-5.tar.gz"

setwd(owd)
unlink(tmp.out)
unlink(tmp.lib)

当然，您可能希望将所有这些都放入一个函数中，最好带有一个选项，以 not 销毁输出目录，这样您就可以根据需要检查.Rout文件和任何.Rout.fail文件。