New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Armadillo 8.100, call .sync() before accessing the internal arrays of the SpMat class #171

Closed
wants to merge 1 commit into
base: master
from

Conversation

Projects
None yet
3 participants
@binxiangni
Contributor

binxiangni commented Aug 31, 2017

I just call .sync() in as<>() and it works. But it seems a little weird that no error comes out when I don't call .sync() in wrap() before reading those internal arrays of the SpMat class.

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Aug 31, 2017

Member

Your tests may be missing the salient fact that we are not on Armadillo 0.8[01]00.* yet.

But it is good that you are proactive and on top of this.

Member

eddelbuettel commented Aug 31, 2017

Your tests may be missing the salient fact that we are not on Armadillo 0.8[01]00.* yet.

But it is good that you are proactive and on top of this.

@binxiangni

This comment has been minimized.

Show comment
Hide comment
@binxiangni

binxiangni Aug 31, 2017

Contributor

So in addition to copying armadillo & armadillo_bits to here and modifying as<>(), what else should I do?

Contributor

binxiangni commented Aug 31, 2017

So in addition to copying armadillo & armadillo_bits to here and modifying as<>(), what else should I do?

@binxiangni

This comment has been minimized.

Show comment
Hide comment
@binxiangni

binxiangni Aug 31, 2017

Contributor

I find that I don't need to call .sync() before arma::access::rwp, but I should call that before arma::access::rw, which possibly means no need to call .sync() before directly reading or writing the pointers. I guess that might explain my problem.

Contributor

binxiangni commented Aug 31, 2017

I find that I don't need to call .sync() before arma::access::rwp, but I should call that before arma::access::rw, which possibly means no need to call .sync() before directly reading or writing the pointers. I guess that might explain my problem.

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Aug 31, 2017

Member

So in addition to copying armadillo & armadillo_bits to here and modifying as<>(), what else should I do?

I think you can wait for more tests until I had a chance to update Armadillo within RcppArmadillo.

Member

eddelbuettel commented Aug 31, 2017

So in addition to copying armadillo & armadillo_bits to here and modifying as<>(), what else should I do?

I think you can wait for more tests until I had a chance to update Armadillo within RcppArmadillo.

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

I am now (at last) at Armadillo 8.100.1. I think I will need the first part of the PR -- as there are some errors -- but there are "issues" with this PR as it covers 80+ files it should not cover (from Armadillo 8.*). Shall we / can we clean that up, or shall I just re-apply the relevant part by hand?

Member

eddelbuettel commented Oct 2, 2017

I am now (at last) at Armadillo 8.100.1. I think I will need the first part of the PR -- as there are some errors -- but there are "issues" with this PR as it covers 80+ files it should not cover (from Armadillo 8.*). Shall we / can we clean that up, or shall I just re-apply the relevant part by hand?

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

(Copying over. Wrote this in #175 by mistake at first.)

This was a very important and timely PR. I had test failures due to SpMat changes, but with your syncing it passes.

Now it so happens that I cannot merge this now, but I will give you full credit in inst/NEWS.Rd. Hope that's ok.

I should have put 8.100.* into a branch earlier. My bad.

Member

eddelbuettel commented Oct 2, 2017

(Copying over. Wrote this in #175 by mistake at first.)

This was a very important and timely PR. I had test failures due to SpMat changes, but with your syncing it passes.

Now it so happens that I cannot merge this now, but I will give you full credit in inst/NEWS.Rd. Hope that's ok.

I should have put 8.100.* into a branch earlier. My bad.

eddelbuettel added a commit that referenced this pull request Oct 2, 2017

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

Ok, I folded this into #176 which is now being tested.

Member

eddelbuettel commented Oct 2, 2017

Ok, I folded this into #176 which is now being tested.

@binxiangni

This comment has been minimized.

Show comment
Hide comment
@binxiangni

binxiangni Oct 2, 2017

Contributor

Sorry for being late to take action. Being busy with midterms. Kind of difficult to catch you up. So what should I do? As I see, you've removed the redundant files, right?

Contributor

binxiangni commented Oct 2, 2017

Sorry for being late to take action. Being busy with midterms. Kind of difficult to catch you up. So what should I do? As I see, you've removed the redundant files, right?

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

No real action item. But if you and @sgsokol can look over the changes I made to RcppArmadilloAs.h in the branch that is behind the #176 PR. It is now running a rev.dep -- we should be good.

Member

eddelbuettel commented Oct 2, 2017

No real action item. But if you and @sgsokol can look over the changes I made to RcppArmadilloAs.h in the branch that is behind the #176 PR. It is now running a rev.dep -- we should be good.

@binxiangni

This comment has been minimized.

Show comment
Hide comment
@binxiangni

binxiangni Oct 2, 2017

Contributor

I checked it and my part should be fine. Thanks! (btw, the PR number above should be #176 )

Contributor

binxiangni commented Oct 2, 2017

I checked it and my part should be fine. Thanks! (btw, the PR number above should be #176 )

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

Ooops. Correcting.

Member

eddelbuettel commented Oct 2, 2017

Ooops. Correcting.

@sgsokol

This comment has been minimized.

Show comment
Hide comment
@sgsokol

sgsokol Oct 2, 2017

Contributor

For me too, it looks OK.

Contributor

sgsokol commented Oct 2, 2017

For me too, it looks OK.

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 2, 2017

Member

Again, sorry to both of for sitting on the PRs in a way that then prevented me from merging them.

Lesson learned, I just should not wait for my next RcppArmadillo release cycle.

Member

eddelbuettel commented Oct 2, 2017

Again, sorry to both of for sitting on the PRs in a way that then prevented me from merging them.

Lesson learned, I just should not wait for my next RcppArmadillo release cycle.

@eddelbuettel

This comment has been minimized.

Show comment
Hide comment
@eddelbuettel

eddelbuettel Oct 3, 2017

Member

The content of this PR is now in master via #176. Again my apologies for not merging this sooner (maybe into a branch with arma 8.*).

Member

eddelbuettel commented Oct 3, 2017

The content of this PR is now in master via #176. Again my apologies for not merging this sooner (maybe into a branch with arma 8.*).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment