use native Armadillo batch constructor for dgt, dtt and dst type conversion (clean PR) #175
Conversation
inst/include/RcppArmadilloAs.h
Outdated
| arma::urowvec tj = mat.slot("j"); | ||
| arma::Col<T> tx = mat.slot("x"); | ||
|
|
||
| res=arma::sp_mat(true, arma::join_cols(ti, tj), tx, nrow, ncol, true, false); |
There was a problem hiding this comment.
Minor bit... But, could you put a space between res = arma::sp_mat()...
|
Thanks for a cleaner new PR, much appreciated. Now, @binxiangni made a good point about the code being able to differentiate between different subtype. Shall we maybe re-add-that feature is a subsequent PR once this one is in? |
|
A valid but dangerous method to make a clean PR is to |
|
@eddelbuettel a commit 4e3ad83 does this differentiation. I have committed it to my master branch (as instructed by github) but I don't know if it will become a part of this PR or of the next one? |
|
@binxiangni I have removed my old fork on github, made a new fork and then cloned it locally. It is a little bit radical way but hopefully, it will work without messing things up. |
If the commit is in the same branch as the previous commits of the PR, then yes. It just grows as you'd expect. Will run a new Travis and just appear here. I'd say try it ... |
|
This was a very important and timely PR. I had test failures due to Now it so happens that I cannot merge this now, but I will give you full credit in I should have put 8.100.* into a branch earlier. My bad. |
|
@eddelbuettel I hope only that it is not my action or non-action that prevents you from an ability to merge. If I can do anything to facilitate the merging this or next time just say it. |
|
My bad -- similar to Binxiang's patch which I sat on for too long, I now merged this "by hand". (I neeed Binxiang's for Armadillo 8.100.1, and with that the file was changed.) But this too is terrific, and very importamt. I added both to the ChangeLog and will update NEWS in a moment before I commit. |
|
Oh, and the earlier comment should gave gone to #171. Ooops. |
|
Ok, I folded this into #176 which is now being tested. |
|
The content of this PR is now in master via #176. Again my apologies for not merging this sooner. |
Rebase didn't help but a fresh new fork fixed the things.