Retirement issue

[rsbivand]

Thanks for your email. It permitted the discovery of a major problem in retiring maptools. The traceback points to car::dataEllipse(), which calls car::showLabels(), and car:::showLabels1() internally, callling maptools::pointLabel(). If maptools imported by car and suggested by Guerry. I'll raise an issue on https://github.com/r-spatial/evolution.

[friendly]

Thanks for tracking this down. I see the problem now.
As I understand this, the problem for maptools does not seem to reside in Guerry, except for the indirect problem with car::showLabels()`

[rsbivand]

Yes, exactly. Without maptools on the library path, the problem arises.

[sdray]

Hi Roger and Michael

thanks for the message. In adespatial, the only occurrence of maptools is the use of pointLabel in one function. I will check if I can remove and use another function. Otherwise, I will try to implement a light version of pointLabel directly in adespatial (I would like to avoid dependency to car if pointLabel move in it).

For Guerry, I found only one occurrence in a outdated vignette that I wrote (MultiSpat.Rnw in folder vignette-old). I think that this folder can be removed as I wrote a new version of this vignette (MultiSpat.Rmd in folder vignettes). Perhaps, we could also move Michael's vignette (guerry-multivariate.Rmd from vignettes-new to vignettes) to clean the structure of the package. Your vignette will then be available on the CRAN page of the package at https://cran.r-project.org/web/packages/Guerry/index.html. Michael, what is your feeling ?

Hence, I think that we can remove now the dependency to maptools in Guerry withous problems (as the function pointLabel is used only in a vignette that is not compiled and no more useful).

Cheers

[friendly]

I will remove the MultiSpat.Rnw from vignettes-old. I got stuck on the guerry-multivariate.Rmd vignette, largely, I recall, b/c I didn't have a biplot function I liked, i.e., one that allowed data ellipses for the observation points when there are groups. I'll take another look.

[friendly]

Working on a new release, to be v.1.8.0.
I think I can now remove Suggests: maptools but will have to test this out. Comments?

[sdray]

Yes ! I think this is sufficient Le 04/10/2022 à 19:59, Michael Friendly a écrit :

…

Working on a new release, to be v.1.8.0. I think I can now remove |Suggests: maptools| but will have to test this out. Comments? — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADKQTDFTGCXB556E3L2VODDWBRWA7ANCNFSM6AAAAAAQ3MLOWU>. You are receiving this because you commented.Message ID: ***@***.***>

-- Stéphane DRAY ***@***.***) Laboratoire BBE-CNRS-UMR-5558, Univ. C. Bernard - Lyon I 43, Bd du 11 Novembre 1918, 69622 Villeurbanne Cedex, France http://pbil.univ-lyon1.fr/members/dray/ GdR EcoStat :https://sites.google.com/site/gdrecostat/

[rsbivand]

I'll deprecate pointLabel() in maptools on R-forge later today and run reverse dependency tests with the development versions of car from R-forge and Guerry from github. Maybe Guerry would need car > a specific version at that stage?

[friendly]

I just removed maptools and ran devtools::check_win_devel(). Result is in: https://win-builder.r-project.org/lN71kf4nW72s/00check.log

I get no problem I can see related to maptools, so I think may be safe.

However, I get an error I don't understand from building the MultiSpat vignette:

Error(s) in re-building vignettes:
--- re-building 'MultiSpat.Rmd' using rmarkdown

Attaching package: 'adegraphics'

The following objects are masked from 'package:ade4':

    kplotsepan.coa, s.arrow, s.class, s.corcircle, s.distri, s.image,
    s.label, s.logo, s.match, s.traject, s.value, table.value,
    triangle.class

Loading required package: spData
To access larger datasets in this package, install the spDataLarge
package with: `install.packages('spDataLarge',
repos='https://nowosad.github.io/drat/', type='source')`
Loading required package: sf
Linking to GEOS 3.9.1, GDAL 3.4.3, PROJ 7.2.1; sf_use_s2() is TRUE

Attaching package: 'spdep'

The following object is masked from 'package:ade4':

    mstree


Attaching package: 'adespatial'

The following object is masked from 'package:ade4':

    multispati

--- finished re-building 'MultiSpat.Rmd'

[sdray]

I think that the problem is related to the other vignette (no ggbiplot package found):

Quitting from lines 42-52 (guerry-multivariate.Rmd)
Error: processing vignette 'guerry-multivariate.Rmd' failed with diagnostics:
there is no package called 'ggbiplot'
--- failed re-building 'guerry-multivariate.Rmd'

[sdray]

Roger, I think that we do not need any dependency to car

[friendly]

I'm removing stuff in my vignette related to two packages which are not on CRAN: ggbiplot and ggradar

[rsbivand]

I cannot check yet because the car maintainers have not yet committed their updates to R-Forge.

[rsbivand]

With car 3.1-1 from R-Forge, the problem from dataEllipse() is resolved.
Guerry-Ex.pdf
John and I also consider that pointLabel() might rather be a candidate for the base graphics package, so I'll ask Paul Murrell. Following an email exchange, we chose car.

[friendly]

Thanks for pursuing this. I've already removed maptools