Skip to content

Latest commit

 

History

History
450 lines (310 loc) · 17.7 KB

cran-comments.md

File metadata and controls

450 lines (310 loc) · 17.7 KB

Version 0.1.0

Test environments

  • local R installation, R 4.0.2
  • ubuntu 16.04 (on travis-ci), R 4.0.2
  • win-builder (devel)

R CMD check results

0 errors | 0 warnings | 1 note

  • This is a new release.

Round 1 (after automatic checks) the 01/12/2020

problem:
Possibly mis-spelled words in DESCRIPTION:
listw (10:23)
reticular (11:52)
spNetwork (8:18)

correction:
These words are not mis-spelled.



**problem**: Found the following (possibly) invalid URLs: URL: http://github.com/JeremyGelb/spNetwork (moved to https://github.com/JeremyGelb/spNetwork) URL: http://github.com/JeremyGelb/spNetwork/issues (moved to https://github.com/JeremyGelb/spNetwork/issues)

correction:
The http were replaced by https



**problem**: Package in Depends/Imports which should probably only be in LinkingTo: 'RcppArmadillo'

correction:
RcppArmadillo is directly used in the package, this note is spurious



**problem**: checking whether package 'spNetwork' can be installed ... WARNING Found the following significant warnings: code.cpp:601:18: warning: suggest parentheses around comparison in operand of '&' [-Wparentheses] code.cpp:712:18: warning: suggest parentheses around comparison in operand of '&' [-Wparentheses] code.cpp:974:20: warning: suggest parentheses around comparison in operand of '&' [-Wparentheses] code.cpp:1083:20: warning: suggest parentheses around comparison in operand of '&' [-Wparentheses]

correction:
The parentheses were added as suggested



**problem**: checking data for non-ASCII characters ... WARNING AREA["Canada - Quebec and Ontario - 75W to 72W"]

correction: This warning is identified for each dataset provided with the package spNetwork. The error is caused by the way SpatialDataFrames from sp package store CRS information. I removed all the datasets from the data file and saved them in a classical GIS format (gpkg) in the folder inst/extdata. I edited all the examples, vignettes and tests to use load the data from this directory. However, this leads to a quite large directory (3.4 MB) and raise a new note.



**problem**: Examples with CPU (user + system) or elapsed time > 10s

correction: all the examples with long example time are not ran anymore during checks (\dontrun{})



### R CMD check results

0 errors | 0 warnings | 3 note

checking CRAN incoming feasibility ... NOTE
Maintainer: 'Jeremy Gelb [email protected]'

New submission
(this is a new submission, no problem here)

checking package dependencies ... NOTE
Package in Depends/Imports which should probably only be in LinkingTo: 'RcppArmadillo'

(as stated above, if RcppArmadillo is only linked, errors occur)

checking installed package size ... NOTE
installed size is 11.1Mb
sub-directories of 1Mb or more:
extdata 3.6Mb
libs 6.4Mb

(as stated above, libs directory is large because of the use of Rcpp. The directory extdata is large too because data used for examples raises warnings when included as .rda files. The datasets are SpatialDataFrames from sp package and they store their Coordinates Reference System with non-ASCII characters. They are thus stored in geopackage files, in extdata directory.)

Round 2 (after first comments) the 07/12/2020

Problem:
please omit the redudnant "The spNetwork package provides functions to perform ". Perhaps simply replace by "Performs"

correction: The DESCRIPTION has been modified as suggested:
The spNetwork package performs spatial analysis on network.

Problem:
Check: Overall checktime, Result: NOTE Overall checktime 28 min > 10 min

mainly from

  • checking re-building of vignette outputs ... [21m] OK

Can this be reduced to few minuted to keep the overall check time < 10 min? For example, use few itarations, toy data and if not possible by reducing the comlexity of the computations, please provide precomputed results for the lengthy parts.

correction: The duration of vignettes building has been reduced by precomputing the results for the two longest vignettes.

The new durations are :
KNetworkFunctions.Rmd : 18 seconds
SpatialWeightMatrices.Rmd : 79 seconds
NKDE.Rmd : 69 seconds
NKDEdetailed.Rmd : 8 seconds

Round 3 (after second comments) the 31/12/2020

" Size of tarball: 37777188 bytes

Overall checktime 16 min > 10 min"

The size of the tarball has been reduced to 3.5 Mo (one data file was not compressed, my bad). The Overall checktime is now only 4min on my laptop.

Round 4 (after third comments, human validation) the 08/01/2021

Problem:
Please omit the redundant "The spNetwork package" from your description.

correction: The description has been modified as requested

Problem: Please always write package names, software names and API (application programming interface) names in single quotes in title and description. e.g: --> 'spNetwork'

correction: As requested, packages names and API are now written with single quotes in the description and the vignettes. The spNetwork name is still in bold in the vignettes.

Problem: if there are references describing the methods in your package, please add these in the description field of your DESCRIPTION file in the form authors (year) doi:... authors (year) arXiv:... authors (year, ISBN:...) or if those are not available: <[https:...]https:...> with no space after 'doi:', 'arXiv:', 'https:' and angle brackets for auto-linking. (If you want to add a title as well please put it in quotes: "Title")

correction: The main references were added in the description as suggested.

Problem: Warning: Unexecutable code in man/cross_kfunctions.mc.Rd: Warning: Unexecutable code in man/kfunctions.mc.Rd: Warning: Unexecutable code in man/lixelize_lines.mc.Rd: Warning: Unexecutable code in man/network_listw.mc.Rd: Warning: Unexecutable code in man/nkde.mc.Rd:

correction: The problem is caused by the use of dontshow for the multicore functions. I removed it. I think it is a good choice to show to the user how to come back to the sequential plan after multiprocessing.

Problem: Please always make sure to reset to user's options(), working directory or par() after you changed it in examples and vignettes and demos. e.g.: inst/doc/NKDEdetailed.R, inst/doc/KNetworkFunctions.R oldpar <- par(mfrow = c(1,2)) ... par(oldpar)

correction: As suggested, the modifications applied to the user's option, par and working directory are now reset after vignettes. These parameters are never modified in examples.

Version 0.1.1

Round 1 (after automatic checks) the 18/01/2021

The package was accepted on CRAN today, but an error occurred on SOLARIS. The error occurs when I set manually the CRS of sp objects in the test and in the vignette NKDEdetailed. I removed the lines specifying the projection (maybe an error with PROJ4 ?). Because the package has already been accepted on CRAN, I have to raise the version number.

Version 0.2.0

submitted the 08/10/2021

This is the result of RCMD check before submission on windows devel version.

-- R CMD check results ------------------------------------------------------ spNetwork 0.2.0 ---- Duration: 8m 21.4s

checking installed package size ... NOTE installed size is 9.5Mb sub-directories of 1Mb or more: doc 2.0Mb extdata 4.0Mb libs 2.8Mb

0 errors √ | 0 warnings √ | 1 note x

github actions were also used to test on: windows-latest, macOS-latest and Ubuntu-20.4

Round 1 (after automatic checks)

3 NOTES were raised:

problem: Possibly misspelled words in DESCRIPTION: isochrones (11:77)

correction: This word is pretty common in spatial analysis. However,I changed it in the description for "reachable area".

problem: Found the following (possibly) invalid URLs: URL: https://github.com/https://jeremygelb.github.io/spNetwork/

correction This error is caused by the use of the package badger. This badge is now created by hand.

problem: Non-standard file/directory found at top level: 'man-roxygen'

correction This folder was added to .Rbuildignore

Version 0.2.1

This version is submitted to correct 2 warnings and one error in CRAN checks on multiple platforms.

submitted the 12/10/2021

problem: Pandoc is required to build R Markdown vignettes but not available. Please make sure it is installed. (on r-release-macos-x86_64 and r-oldrel-macos-x86_64 )

correction: This was caused by a modified header for the new vignette. This has been fixed.

problem: Error: C++17 standard requested but CXX17 is not defined

  • removing ‘/home/ripley/R/Lib32/spNetwork’

correction: It seems that SOLARIS x86 does not support C++17. I find a solution in the package rcppsimdjson facing the same issue. It uses a configure file checking that c++17 is available and switch to c++11 otherwise. A warning message is also printed for the user. I hope it will work because I can not test it locally nor on rhub.

round 2 (after manual check by Uwe Ligges)

problem:

""" Thanks, we see:

with gcc-UBSAN:

spNetwork.Rcheck/spNetwork-Ex.Rout:/data/gannet/ripley/R/test-4.2/RcppArmadillo/include/armadillo_bits/subview_meat.hpp:1433:54: runtime error: reference binding to null pointer of type 'const double' (many times)

spNetwork.Rcheck/spNetwork-Ex.Rout:/data/gannet/ripley/R/test-4.2/RcppArmadillo/include/armadillo_bits/access.hpp:28:100: runtime error: reference binding to null pointer of type 'double' spNetwork.Rcheck/spNetwork-Ex.Rout:/data/gannet/ripley/R/test-4.2/RcppArmadillo/include/armadillo_bits/subview_meat.hpp:1433:54: runtime error: reference binding to null pointer of type 'const double'

with clang-UBSAN:

spNetwork.Rcheck/build_vignettes.log:/data/gannet/ripley/R/test-clang/RcppArmadillo/include/armadillo_bits/subview_meat.hpp:1433:23: runtime error: reference binding to null pointer of type 'const double' 4 times """

correction: After inspection, it seems that these errors occured when building the vignette NetworkBuilding. More specifically, it seems linked to the c++ function spNetwork_split_lines_at_points_cpp. The error was probably caused by an edge case were no lines need to be split. In that case, we try to reach elements in an empty (nrow = 0) arma matrix. This has been fixed by catching this edge case before calling the c++ function. When the lines do not need to be split, the input lines are directly returned. It is interesting that the error is not raised when advanced testing (sanitizer) is not used (and the results are valid according to the unit tests).

NOTE: I tried to replicate the error by installing docker and following the instructions here: http://dirk.eddelbuettel.com/blog/2015/01/18/. However, I was not able to perform the test because of an error in the lattice package generated when sanitizer tests are applied.

round 3 (after automatic check)

problem:

  • checking examples ... [106s] NOTE Examples with CPU (user + system) or elapsed time > 10s user system elapsed split_lines_at_vertex 13.03 0.24 13.41 simplify_network 9.78 0.07 10.01

correction: Applied \donttest{} to both functions.

Round 1 13/10/2021 (after Ligges check)

problem: Found the following (possibly) invalid URLs: URL: https://codecov.io/gh/JeremyGelb/spNetwork?branch=master (moved to https://app.codecov.io/gh/JeremyGelb/spNetwork?branch=master) From: README.md Status: 301 Message: Moved Permanently

correction:

The URL was replaced as suggested in the Readme

Version 0.4.2

This version is submitted to correct 2 errors raised because of the new compiler gcc12 and the vignette using rgl. I did not had enough time to do the corrections and spNetwork was removed from CRAN.

submitted the 15/04/2022

round 2 (after Gregor's check)

Problem: Size of tarball: 5130234 bytes A CRAN package should not be larger than 5 MB. Please reduce the size.

Correction: Some files used to build the package website were not added to buildignore. This has been corrected and the package weight is now 4602 Ko

Problem: Please write TRUE and FALSE instead of T and F. (Please don't use 'T' or 'F' as vector names.), e.g.: man/split_by_grid.mc.Rd: split_by_grid.mc(grid, samples, events, lines, bw, tol, digits, split_all = T)

Correction: I looked up for the abreviations T and F and changed them for TRUE and FALSE

Problem: Please add \value to .Rd files regarding exported methods and explain the functions results in the documentation. Please write about the structure of the output (class) and also what the output means. (If a function does not return a value, please document that too, e.g. \value{No return value, called for side effects} or similar) Missing Rd-tags: bw_cv_likelihood_calc_tkde.Rd: \value gfunc_cpp.Rd: \value kfunc_cpp.Rd: \value tkde.Rd: \value tnkde_worker_bw_sel.Rd: \value tnkde.mc.Rd: \value tnkde.Rd: \value

correction: For all the functions above, a @return entry has been added in the Roxygen documentation

Version 0.4.3

submitted the 20/04/2022

Problem:

noLD error in one test on macOS, likely caused by a rounding issue. The test has been rewritten to be more tolerant.

Version 0.4.3.1

submitted the 20/04/2022

Problem: Warning in CPL_read_ogr(dsn, layer, query, as.character(options), quiet, : GDAL Error 1: unable to open database file: this file is a WAL-enabled database. It cannot be opened because it is presumably read-only or in a read-only directory. Error: Cannot open "/home/hornik/tmp/R.check/r-devel-gcc/Work/build/Packages/spNetwork/extdata/events.gpkg"; The source could be corrupt or not supported. See st_drivers() for a list of supported formats. Execution halted

The check problems on the Debian systems are caused by attempts to write to the user library to which all packages get installed before checking (and which now is remounted read-only for checking).

correction: All the data provided as gpkg files are now provided in .rda files in the data folder. The CRS has been changed to avoid error with non-ASCII characters.

Version 0.4.3.2

submitted the 13/05/2022

Problem: SearchTrees package will be soon removed from CRAN.

correction: We changed the functions using it to remove the dependency.

round 2 (after automatic checks)

Problem: 1 NOTE: Found if() conditions comparing class() to string:

correction: This has been replaced as suggested by the inherits function

Version 0.4.4.2

submitted the 20/05/2024

Problem: c++ compilation warning on gcc-14 (GCC) 14.1.0 x86_64-pc-linux-gnu

Found the following significant warnings: bw_selection_tnkde.cpp:760:85: warning: ISO C++ forbids compound-literals [-Wpedantic] bw_selection_tnkde.cpp:765:83: warning: ISO C++ forbids compound-literals [-Wpedantic] nkde_continuous.cpp:127:151: warning: ISO C++ forbids compound-literals [-Wpedantic] nkde_continuous.cpp:134:145: warning: ISO C++ forbids compound-literals [-Wpedantic] nkde_discontinuous.cpp:112:81: warning: ISO C++ forbids compound-literals [-Wpedantic]

correction: The c++ code has been modified to avoid the compound-literals

Version 0.4.4.3

submitted the 23/05/2024

Problem: complex UBSAN warning raised in tests:

/data/gannet/ripley/R/test-dev/RcppArmadillo/include/armadillo_bits/access.hpp:28:100: runtime error: reference binding to null pointer of type 'double' #0 0x7f37bbda38bf in double& arma::access::rw(double const&) /data/gannet/ripley/R/test-dev/RcppArmadillo/include/armadillo_bits/access.hpp:28 #1 0x7f37bbda38bf in arma::Mat::colptr(unsigned int) /data/gannet/ripley/R/test-dev/RcppArmadillo/include/armadillo_bits/Mat_meat.hpp:6759 #2 0x7f37bbda38bf in arma::Mat<arma::Mat::elem_type> arma::subview_each1_aux::operator_minus<arma::Mat, 1u, arma::subview >(arma::subview_each1<arma::Mat, 1u> const&, arma::Base<arma::Mat::elem_type, arma::subview > const&) /data/gannet/ripley/R/test-dev/RcppArmadillo/include/armadillo_bits/subview_each_meat.hpp:750 #3 0x7f37bbdb8ab5 in arma::Mat<arma::Mat::elem_type> arma::operator-<arma::Mat, 1u, arma::subview >(arma::subview_each1<arma::Mat, 1u> const&, arma::Base<arma::Mat::elem_type, arma::subview > const&) /data/gannet/ripley/R/test-dev/RcppArmadillo/include/armadillo_bits/operator_minus.hpp:517 #4 0x7f37bbdb8ab5 in esd_kernel_rcpp_arma_sparse(arma::Col (*)(arma::Col, double), arma::SpMat&, Rcpp::Vector<19, Rcpp::PreserveStorage>&, int, double, arma::Col&, arma::Col&, arma::Mat&, arma::Mat&, int, int) /data/gannet/ripley/R/packages/tests-gcc-SAN/spNetwork/src/nkde_discontinuous.cpp:100 #5 0x7f37bbdbfc88 in discontinuous_nkde_cpp_arma_sparse(Rcpp::Vector<19, Rcpp::PreserveStorage>, Rcpp::Vector<14, Rcpp::PreserveStorage>, Rcpp::Vector<14, Rcpp::PreserveStorage>, Rcpp::DataFrame_ImplRcpp::PreserveStorage, Rcpp::Vector<14, Rcpp::PreserveStorage>, std::__cxx11::basic_string<char, std::char_traits, std::allocator >, Rcpp::DataFrame_ImplRcpp::PreserveStorage, Rcpp::DataFrame_ImplRcpp::PreserveStorage, int, bool, std::__cxx11::basic_string<char, std::char_traits, std::allocator >) /data/gannet/ripley/R/packages/tests-gcc-SAN/spNetwork/src/nkde_discontinuous.cpp:308 #6 0x7f37bbac2766 in _spNetwork_discontinuous_nkde_cpp_arma_sparse /data/gannet/ripley/R/packages/tests-gcc-SAN/spNetwork/src/RcppExports.cpp:892

correction: The C++ has been edited for the targeted function. I have not been able to reproduce the error but I found some guidance about memory overflow when comparing integers and unsigned integers.