Building with intel compilers

[gareth-d-ga]

When trying to install s2 on a machine that uses the Intel compilers, I get this error message:

icpc -std=gnu++11 -I"/apps/R/4.2.1/lib64/R/include" -DNDEBUG -I../src -DSTRICT_R_HEADERS -I'/home/547/gxd547/R/x86_64-pc-linux-gnu-library/4.2/Rcpp/include' -I'/home/547/gxd547/R/x86_64-pc-linux-gnu-library/4.2/wk/include' -I/usr/local/include  -DIS_LITTLE_ENDIAN -pthread -fpic  -g -O2  -c absl/strings/cord.cc -o absl/strings/cord.o
In file included from ../src/absl/strings/cord.h(84),
                 from absl/strings/cord.cc(15):
../src/absl/strings/internal/string_constant.h(43): error: expression must have a constant value
    static_assert(value.empty() || 2 * value[0] != 1,
                  ^
../src/absl/strings/internal/string_constant.h(43): note: expression cannot be interpreted
    static_assert(value.empty() || 2 * value[0] != 1,
                                ^

compilation aborted for absl/strings/cord.cc (code 2)
make: *** [/apps/R/4.2.1/lib64/R/etc/Makeconf:175: absl/strings/cord.o] Error 2
ERROR: compilation failed for package ‘s2’

I'm not a c++ programmer, but was wondering if you could have a look to see if the code is valid c++ (or else, it's a compiler issue)?

You might ask why I don't just use gcc. On this machine, it will be difficult as I would need to build a lot of dependencies myself, which have been pre-built by the administrators with Intel. While I only recently started using sf and dependencies, for many years previously I could successfully used R spatial packages with these compilers.

My sessionInfo() is copied below.

Thanks,
Gareth

> sessionInfo()
R version 4.2.1 (2022-06-23)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Rocky Linux 8.6 (Green Obsidian)

Matrix products: default
BLAS/LAPACK: /apps/intel-oneapi/mkl/2022.1.0/lib/intel64/libmkl_intel_lp64.so.2

locale:
 [1] LC_CTYPE=en_AU.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_AU.UTF-8        LC_COLLATE=en_AU.UTF-8    
 [5] LC_MONETARY=en_AU.UTF-8    LC_MESSAGES=en_AU.UTF-8   
 [7] LC_PAPER=en_AU.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C       

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

loaded via a namespace (and not attached):
[1] compiler_4.2.1 tools_4.2.1    tcltk_4.2.1   

[gareth-d-ga]

Confirming that if I comment out the static_assert statement flagged above (lines 43-44 of s2/src/absl/strings/internal/string_constant.h), then I can build s2 OK.

I am using the following software:

[XXXXX@gadi-login-05 s2]$ module list
Currently Loaded Modulefiles:
 1) intel-compiler/2021.6.0   3) netcdf/4.7.3(default)   5) R/4.2.1              7) python3/3.10.4   9) udunits/2.2.26  
 2) gcc/system                4) geos/3.8.0              6) intel-mkl/2021.4.0   8) gdal/3.5.0      10) proj/6.2.1      

[paleolimbot]

Thanks for tracking this down! I'm happy to include these changes in our vendored Abseil.