Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upIncorrectly specified init.c routine registration? #274
Comments
|
Didn't mean to close. Here is the corrected version. I'd PR but I don't think I can push to github from work.
|
|
Well it builds at CRAN and on win-builder so my first reaction would be to blame local problems... |
|
If you changed |
|
Ooh good point. I'm just seeing that your I deleted the Did |
|
And |
|
Ah, now things are making sense. Yes, removing Running it now locally to ensure it builds, but I think this might be a worthwhile change as the next time someone runs |
|
Things look good after removing And I believe here is that Rcpp commit that adds the leading underscore as an FYI. |
|
A few other repos had similar changes of removing |
|
It's okay. Thanks for being responsive. If you give me a few minutes I think I can actually submit a PR. |
Closes #274. Remove src/init.c and use the newer version of compileAt…
I've been trying to build on Windows 7 with R 3.4.1 and keep running into:
After much suffering, I found that in your
init.cfile you are registering routines using:Rblpapi_authenticate_Implbut Rcpp exports these as_Rblpapi_authenticate_Impl. Don't you need to be consistent for this to work correctly?I went ahead and cloned and prefixed everything with underscores and it seems to build correctly now.
(Keep in mind I know a decent amount about R, but essentially nothing about C/C++ so this is the result of speculation + reading the R Extensions section on this)
Extra side note) I looked at how
dplyrdoes this and they define their call entries and register the routines in theirRcppExports.cppfile. They also don't have thatextern SEXP ...piece. Not sure if that part is necessary or not, as again I essentially know nothing in this space.https://github.com/tidyverse/dplyr/blob/master/src/RcppExports.cpp