Skip to content

Latest commit

 

History

History
119 lines (78 loc) · 3.14 KB

cran-comments.md

File metadata and controls

119 lines (78 loc) · 3.14 KB

landscapemetrics 1.5.3

Minor bug fixes

landscapemetrics 1.5.2

Minor bug fixes

landscapemetrics 1.5.1

Small improvements package CI and bug fixes

landscapemetrics 1.5.0

Better handling of input data, new functions and terra support

landscapemetrics 1.4.7

Adding new function

landscapemetrics 1.4.6

Minor improvements of existing functions

landscapemetrics 1.4.5

Minor improvements, bug fixes and new functions

landscapemetrics 1.4.4

Minor improvements and bug fixes

landscapemetrics 1.4.3

Improvements of existing functions

landscapemetrics 1.4.2

Minor improvements and bug fixes

landscapemetrics 1.4.1

Major bugfixes

landscapemetrics 1.4

Improvements of existing functions and make sure all CRAN checks run

landscapemetrics 1.3.2

Make sure all CRAN checks run

Update Version: 1.3.1

Minor bugfixes

Update Version: 1.3

Minor bugfixes

Update Version: 1.2.2

Small bugfixes and improvements of existing functions

Update Version: 1.2.1

Small bugfixes

Update Version: 1.2

Minor bugfixes and improvments of existing functions

Update Version: 1.1

Minor bugfixes and improvments of existing functions

Update Version: 1.0

Minor bugfixes, less dependencies and new function

Update Version: 0.3.1

Minor bugfixes

Update Version: 0.3

Minor bugfixes, better memory allocation and new functions

Update Version: 0.2

Minor bugfixes and new functions

Update Version: 0.1.1

a) The Solaris installation error is described in §1.6.4 of 'Writing R Extensions': please study it.

  • Done as suggested

b) You use isFALSE from R 3.5.x despite claiming Depends: R (≥ 3.1): you cannot have actually tested that.

  • Done as suggested

c) There are problems with the tests showing in the 'Additional issues' (shortly if not already).
The way you write your tests, the reports such as

── 1. Failure: lsm_c_lsi results are equal to fragstats (@test-lsm-c-lsi.R#7) ─ all(...) isn't true.

── 2. Failure: lsm_c_split results are equal to fragstats (@test-lsm-c-split.R#7 all(...) isn't true.

── 3. Failure: lsm_p_enn results are comparable to fragstats (@test-lsm-p-enn.R# all(...) isn't true.

are completely useless. But testing numerical results by rounding rather than using all.equal is bad practice (and it is not you who is paying the price!).

You could use something like

all.equal(sort(fragstats_class_landscape_lsi), sort(landscapemetrics_class_landscape_lsi$value)) [1] "Mean relative difference: 4.816766e-06"

with a suitable tolerance.

  • The software we compare our results to (FRAGSTATS) has only a precision of 4 digits itself. We discussed using relative differences but decided that we do not want to show that our results are equal to the FRAGSTATS results within a tolerance, but rather that our results would be exactly the same assuming the same precision. Therefore we decided to round our results for the tests.

Test environments

  • Windows 10, R 3.5.2
  • ubuntu 18.04, R 3.5.2
  • macOS Mojave, R 3.5.2
  • Windows 10, R (unstable) (2019-01-18 r75994)
  • win-builder (release and devel)

R CMD check results

0 errors | 0 warnings | 0 note

Reverse dependencies

There are currently no reverse dependencies.