LinkingTo secondary LinkingTo packages about mapdeck HOT 11 CLOSED

Just noticed you added BH to LinkingTo in the latest commit.....just tested this again, dev version now installs cleanly on both Mac and PC for me 👍

I'll leave this open for now, but if you're planning on keeping the BH addition edit, then obv feel free to close it.

from mapdeck.

Forgot to mention that this is after installing the dev versions of googlePolylines and jsonify, per instructions here.

Also, session info:

library(googlePolylines)
library(jsonify)
sessionInfo()

R version 3.5.0 (2018-04-23)
Platform: x86_64-apple-darwin15.6.0 (64-bit)
Running under: macOS Sierra 10.12.6

Matrix products: default
BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/3.5/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] jsonify_0.1.0001         googlePolylines_0.7.1004

loaded via a namespace (and not attached):
 [1] rapidjsonr_1.1        httr_1.3.1            compiler_3.5.0        R6_2.3.0              tools_3.5.0          
 [6] withr_2.1.2           curl_3.2              yaml_2.2.0            Rcpp_0.12.19.2        memoise_1.1.0        
[11] knitr_1.20            git2r_0.23.0          colourvalues_0.1.1001 digest_0.6.17         devtools_1.13.6 

from mapdeck.

I had the same issue on Ubuntu 16.04.

from mapdeck.

I just ran these three lines without issue.

devtools::install_github("SymbolixAU/jsonify")
devtools::install_github("SymbolixAU/googlePolylines")
devtools::install_github("SymbolixAU/mapdeck")

However, I also had to add rapidjsonr to LinkingTo, even though mapdeck doesn't use it directly (it uses it through jsonify). But I haven't had issues with BH.
And I also couldn't find another package which links to a package which links to another to another, nor see in any documentation the recommended approach to this.

from mapdeck.

Hmmm, yeah it feels like this should be fixable with some Makevars flags, but I'm not 100% sure, and am not an expert with that stuff.

from mapdeck.

Yes, I'll leave it open until I know the reasons / answer. I've changed the title too if that's OK?

from mapdeck.

Cool, gotcha.

Yep no problem at all with the title change :)

from mapdeck.

Quick update on this.....

I got confirmation that RcppParallel is indeed not actually using boost (see here), so can't use that as an example.

I did find a real world example though...package promises links to Rcpp and later. Package later links to Rcpp and BH.

The API to link to later is set up more like a C library API though....later features an API header file in inst/include, and then uses func R_RegisterCCallable() in an init.c file to make the func execLaterNative() available to linking packages. Their set up is kind of unusual though (see here).

I've worked on the C API for the stringdist package (which uses C functions and does NOT use Rcpp), the way later is set up is somewhat similar....we have an API header file in inst/include that "exports" specific C functions via R_GetCCallable(), then uses R_RegisterCCallable() in the src/init file to expose those functions to linking packages. I'm linking to stringdist in my package refinr. We used the xts C API when designing the stringdist C API.

from mapdeck.

Thanks for looking into this. All this is a bit over my head at the moment, but I should be able to start digging a bit deeper in a couple of weeks after implementing a few other features I've got on my TODO list.

Although, part of me is thinking if it aint broke, and it gets accepted on CRAN, do we need to do anything...

from mapdeck.

No problem :)

Yeah if adding BH to LinkingTo makes it all "just work", I think I would do that if I were you, especially if the alternative is to completely re-write the API for colourvalues.

from mapdeck.

Closing because "it works" at the moment, and I'm ok with this. If it rears its head again down the line we may have to find a proper solution.

from mapdeck.