Open anujkhare opened 5 years ago
just need to make sure it passes R CMD check --as-cran with no errors/warnings, which is what Travis checks. I would suggest a first CRAN submission once we get that done.
after that it would be a good idea to try cv.iregnet on a variety of data sets, perhaps showing these as examples in a vignette?
The travis build does pass. Vignette is partially complete, will work on it once the optimization part is done. Here's the vignette: https://theadityasam.github.io/iregvignette/iregnet.html
Version 0.1.0.0 was submitted to CRAN on 21st August, 2019 and failed on the following points: 1) Produces failed to converge warning (@test_NAN.R#39) 2) On a windows build, ISO C++ forbids variable length array ‘status’ [-Wvla] These are minor issues and will be fixed in the version 1.0.0 soon enough. The next version will also contain the vignette and will be a complete package.
great. to make sure the checks pass on all CRAN machines you should try rhub::check_for_cran()
On Mon, Aug 26, 2019 at 8:32 AM Aditya Samantaray notifications@github.com wrote:
Version 0.1.0.0 was submitted to CRAN on 21st August, 2019 and failed on the following points:
- Produces failed to converge warning (@test_NAN.R#39)
- On a windows build, ISO C++ forbids variable length array ‘status’ [-Wvla] These are minor issues and will be fixed in the version 1.0.0 soon enough. The next version will also contain the vignette and will be a complete package.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/anujkhare/iregnet/issues/67?email_source=notifications&email_token=AAHDX4TBIH3JYMH32QNOMLTQGPZQLA5CNFSM4H33KXD2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5EXNUY#issuecomment-524908243, or mute the thread https://github.com/notifications/unsubscribe-auth/AAHDX4TMQR4QFF463JW3GN3QGPZQLANCNFSM4H33KXDQ .
Okay
Hey @tdhock , one of the warnings from CRAN was that ISO C++ 1998 standard doesn't support long long. What to do in this case? Should I specify the compiler?
why is long long needed? can you just use an int instead? I don't think it is a good idea to depend on specific compilers.
Hey @tdhock, this is in reference to replacing long long with int. The size of int is itself platform dependent being 2 bytes in some sytems and 4 bytes in the rest. Also, we've used long long to store the matrix sizes which might be large enough to cause an overflow in int, that then needs to be handled. Should I specify the compiler options in the Makevars? Will that fix the error?
matrix sizes are typically not so big. please change to int
@tdhock Besides versioning the package and making sure that R CMD checks pass, is there anything else that we need to do?
In general, we can do a lot more testing. @theadityasam: there are a bunch of datasets that Toby mentioned on this page. It'd be a good exercise for us to fit models on these using cv.iregnet.