GSW-R should not "Depend" on the 'testthat' package

[dankelley]

As pointed out by @paleolimbot, this dependency can cause problems for some users. And, apart from the test suite (which is handled without Depending on testthat), there is no real need for this, because documentation entries of the form

expect_equal(A, B)

are just as clear to the user if written as

all.equal(A, B)

[dankelley]

I've started this work. I also see the following, which I'll clean up.

Found the following (possibly) invalid URLs:

• [x] URL: http://dankelley.github.com/oce From: inst/doc/gsw.html Status: 404 Message: Not Found
• [x] URL: http://dankelley.github.io/oce (moved to https://dankelley.github.io/oce/) From: man/gsw.Rd Status: 200 Message: OK
• [x] URL: http://www.teos-10.org/pubs/gsw/html/gsw_enthalpy_ct_exact.html (moved to http://www.teos-10.org/pubs/gsw/html/gsw_enthalpy_CT_exact.html) From: man/gsw_enthalpy_CT_exact.Rd Status: 200 Message: OK
• [x] URL: http://www.teos-10.org/pubs/gsw/html/gsw_saar.html (moved to http://www.teos-10.org/pubs/gsw/html/gsw_SAAR.html) From: man/gsw_SAAR.Rd Status: 200 Message: OK
• [x] URL: http://www.tos.org/oceanography/article/history-of-the-equation-of-state-of-seawater (moved to https://www.tos.org/oceanography/article/history-of-the-equation-of-state-of-seawater) From: inst/doc/gsw.html Status: 200 Message: OK

[dankelley]

I just checked to see whether GSW-C has been updated since the last release of GSW-R, and it seems not, according to the Changelog at https://github.com/TEOS-10/GSW-C (the README file).

However, I think that section must not be maintained, since there have been commits to GSW-C since 2017-02-23, the last date listed there.

Tomorrow morning, I will synch GSW-R up with the most recent GSW-C. And then, assuming all tests pass (my guess is that I'll need to tweak some tests, given the commit notes I see in GSW-C), I'll let GSW-R sit for a week or so, and make a calendar item for next week to do a CRAN release ... unless my coauthor, @richardsc, objects.

[dankelley]

Hm, I see quite a few changes in the key file

diff gsw_saar.c ../GSW-R/src/gsw_saar.c | wc
  419    1537   14401

I will need to check whether the gsw_data.c file is changed a lot, too, but my guess is that the code is changing faster than the data.

[dankelley]

I'll be grabbing the data from the fortran-created netcdf file; see https://github.com/TEOS-10/GSW-R/blob/29c1f62fc7d780f25d1e2071d627c492f90a01e4/developer/create_data/create_saar.R#L2

[dankelley]

The fortran data file (https://github.com/TEOS-10/GSW-Fortran/blob/master/test/gsw_data_v3_0.nc) is 4 years old so, so we are probably up-to-date from the last CRAN release. Even so, I'll use the new version, and I'll insert comments in the R code (and the package docs) stating the md5 hash of this file.

[dankelley]

The .rda I build up from the fortran file is not the same as before (different size, different md5) but that might just be because R builds up rda files differently now.

[dankelley]

All tests still work with the updated data file. I updated the dependency to R >= 3.5.0 because, well, it's long-past time, and it is in this version that the new form of .rda storage was settled upon.

[dankelley]

@efiring -- mentioning you here, so you'll know that GSW-R is going through an update process. No action is required on your part, but I thought you might like to know, since you're the one who most recently updated GSW-C, upon which GSW-R depends. Hope all is well in your locale. Things are great here, as we are coming out of covid. -- Dan.

[efiring]

Dan, thanks for the notification. I hope none of the C changes since your last update cause trouble for you.

[paleolimbot]

A few things that might help that you can feel free to disregard:

• It's possible to mostly automate the updating of the C library (see here), which may help you stay up to date with future releases.
• It looks like you're using macros to generate the wrappers...you might consider generating human-readable C code for the easy-to-wrap functions, which will probably fail more predictably and with better error messages when it does. In my quest to understand GSW I managed to do this for all the functions without pointer args automatically (see here). Likely outside the scope of the changes here but you're certainly welcome to use any of the code in that repo.

Let me know if there's anything I can do! I feel somewhat responsible having suggested the change that it generated quite a bit of work.

[dankelley]

Done in "develop" branch, commit c9ee919e10d8009296b21b2e5a9c760f27cf40c1