Some questions motivated by this discussion. From the CRAN policy page: "Checking the package should take as little CPU time as possible, as the CRAN check farm is a very limited resource and there are thousands of packages. Long-running tests and vignette code can be made optional for checking, but do ensure that the checks that are left do exercise all the features of the package." Is there a further document that elucidates more on this? What is "little CPU time". Is there a documented variable that I can check and then reduce the test set for CRAN? Duncan mentioned one in the dicussion, but I'll end up forgetting the details. A static reference would help. "Test all the features". The test directory for survival has 75 files and still doesn't test everything: I'm about to add a new one today in response to a bug report. How do I adjucate between "little time" and "test all"? Footnote: the manual page for R CMD check (?check) has the line "Many of the checks in R CMD check can be turned off or on by environment variables: see Chapter 6 of the ?R Internals? manual". The reference should be chapter 7. That chapter does document the use of _R_CHECK_TIMINGS_=10. But it seems dangerous to use this as an indirect test, ie. "if timings is 10 then this must be CRAN running". My biggest time offender is in the vignettes. After multiple readings of the docs I still can't quite figure out how to specify - pdf files that should be in the vignettes index, but are not .Rnw source - how to tell R which ones to redo, and which to just accept the pdf - per the policy line "all source for pdf should be available", non-inclusion of the Rnw file isn't an answer. Again, is there another document I'm missing? Terry T.
On Wed, Sep 5, 2012 at 12:08 PM, Terry Therneau <therneau at mayo.edu> wrote:> My biggest time offender is in the vignettes. After multiple readings of > the docs I still can't quite figure out how to specify > - pdf files that should be in the vignettes index, but are not .Rnw > source > - how to tell R which ones to redo, and which to just accept the pdf > - per the policy line "all source for pdf should be available", > non-inclusion of the Rnw file isn't an answer. > Again, is there another document I'm missing?If you have multiple Rnw files producing pdfs, and you only want some of them to be run, you do the following. Say we have big.Rnw (takes a long time to run) small.Rnw (is quick) You run yourself big.Rnw and produce big.pdf. Then you put small.Rnw in /vignettes big.Rnw and big.pdf in inst/doc That way, big.Rnw will not get run. And the package installation will contain both Rnw files and pdf files for both vignettes. Note that this only works if you have both a quick and a long vignette. If you only have big.Rnw and big.pdf in inst/doc and nothing in vignettes, R would still run Sweave on big.Rnw. I agree the section on this in R-exts is a bit opaque: "By default R CMD build will run Sweave on all files in Sweave format in vignettes, or if that does not exist, inst/doc (but not in sub-directories)." (R-exts 1.4) Kasper