search

kaskr / adcomp

AD computation with Template Model Builder (TMB)
Other
176 stars 80 forks source link

build failure with development version of Matrix (1.5.4): NEW_INTEGER macro not found #373

Closed bbolker closed 1 year ago

bbolker commented 1 year ago

Installing from source with the latest version of R (irrelevant) and the development version of Matrix breaks.

I'm not exactly sure what happened but there are comments in the Matrix package SVN repository about moving from "Rdefines.h" to "Rinternals.h"; the NEW_INTEGER macro is defined in Rdefines.h, and the comments in Rdefines.h say

Rdefines.h macros for an S-like interface to the above (no longer maintained)

I think Matrix used to import this but doesn't any more. A fix might be to

#define NEW_INTEGER(n)      Rf_allocVector(INTSXP,n)

but I haven't tried it. (For now I reverted Matrix to the earlier [CRAN] version.) (Alternatively could explicitly #include <Rdefines.h>, but maybe that's suboptimal since Matrix is moving away from it?)

sessionInfo():

R Under development (unstable) (2022-12-31 r83530)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Pop!_OS 22.04 LTS

Matrix products: default
BLAS:   /usr/local/lib/R/lib/libRblas.so 
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3;  LAPACK version 3.10.0

<locale/time zone info redacted>      

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

other attached packages:
[1] Matrix_1.5-4

loaded via a namespace (and not attached):
[1] compiler_4.3.0  bspm_0.3.10     grid_4.3.0      lattice_0.20-45
bolker@pop-os:~/R/pkgs/adcomp$ make install
make build-package
make[1]: Entering directory '/home/bolker/Documents/R/pkgs/adcomp'
R CMD build --resave-data=no TMB
Loading required package: utils
Tracing function "install.packages" in package "utils"
* checking for file ‘TMB/DESCRIPTION’ ... OK
* preparing ‘TMB’:
* checking DESCRIPTION meta-information ... OK
* cleaning src
* checking for LF line-endings in source and make files and shell scripts
* checking for empty or unneeded directories
* building ‘TMB_1.9.1.tar.gz’

make[1]: Leaving directory '/home/bolker/Documents/R/pkgs/adcomp'
R CMD INSTALL --preclean TMB_1.9.1.tar.gz
Loading required package: utils
Tracing function "install.packages" in package "utils"
* installing to library ‘/usr/local/lib/R/site-library’
* installing *source* package ‘TMB’ ...
** using staged installation
** libs
using C compiler: ‘gcc (Ubuntu 11.3.0-1ubuntu1~22.04) 11.3.0’
make[1]: Entering directory '/tmp/RtmpNffoEw/R.INSTALL137d4489d8994/TMB/src'
make[1]: Leaving directory '/tmp/RtmpNffoEw/R.INSTALL137d4489d8994/TMB/src'
make[1]: Entering directory '/tmp/RtmpNffoEw/R.INSTALL137d4489d8994/TMB/src'
gcc -I"/usr/local/lib/R/include" -DNDEBUG  -I'/usr/local/lib/R/site-library/Matrix/include' -I/usr/local/include   -fopenmp -fpic  -g -O2  -c external_metis.c -o external_metis.o
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from external_metis.c:7:
external_metis.c: In function ‘have_tmb_symbolic’:
external_metis.c:104:17: warning: implicit declaration of function ‘NEW_INTEGER’; did you mean ‘NA_INTEGER’? [-Wimplicit-function-declaration]
  104 |   PROTECT(ans = NEW_INTEGER(1));
      |                 ^~~~~~~~~~~
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
external_metis.c:104:15: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  104 |   PROTECT(ans = NEW_INTEGER(1));
      |               ^
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
gcc -I"/usr/local/lib/R/include" -DNDEBUG  -I'/usr/local/lib/R/site-library/Matrix/include' -I/usr/local/include   -fopenmp -fpic  -g -O2  -c init.c -o init.o
gcc -I"/usr/local/lib/R/include" -DNDEBUG  -I'/usr/local/lib/R/site-library/Matrix/include' -I/usr/local/include   -fopenmp -fpic  -g -O2  -c local_stubs.c -o local_stubs.o
gcc -I"/usr/local/lib/R/include" -DNDEBUG  -I'/usr/local/lib/R/site-library/Matrix/include' -I/usr/local/include   -fopenmp -fpic  -g -O2  -c solve_subset.c -o solve_subset.o
solve_subset.c: In function ‘tmb_as_cholmod_factor3’:
solve_subset.c:81:23: warning: implicit declaration of function ‘GET_SLOT’ [-Wimplicit-function-declaration]
   81 |   int *type = INTEGER(GET_SLOT(x, install("type")));
      |                       ^~~~~~~~
solve_subset.c:81:23: warning: passing argument 1 of ‘INTEGER’ makes pointer from integer without a cast [-Wint-conversion]
   81 |   int *type = INTEGER(GET_SLOT(x, install("type")));
      |                       ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
      |                       |
      |                       int
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from solve_subset.c:53:
/usr/local/lib/R/include/Rinternals.h:272:22: note: expected ‘SEXP’ {aka ‘struct SEXPREC *’} but argument is of type ‘int’
  272 | int  *(INTEGER)(SEXP x);
      |                 ~~~~~^
solve_subset.c:93:7: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
   93 |   tmp = GET_SLOT(x, Matrix_permSym);
      |       ^
solve_subset.c:95:27: warning: passing argument 1 of ‘INTEGER’ makes pointer from integer without a cast [-Wint-conversion]
   95 |   ans->ColCount = INTEGER(GET_SLOT(x, install("colcount")));
      |                           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      |                           |
      |                           int
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from solve_subset.c:53:
/usr/local/lib/R/include/Rinternals.h:272:22: note: expected ‘SEXP’ {aka ‘struct SEXPREC *’} but argument is of type ‘int’
  272 | int  *(INTEGER)(SEXP x);
      |                 ~~~~~^
solve_subset.c:98:7: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
   98 |   tmp = GET_SLOT(x, Matrix_xSym);
      |       ^
solve_subset.c:104:9: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  104 |     tmp = GET_SLOT(x, install("super"));
      |         ^
solve_subset.c:106:9: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  106 |     tmp = GET_SLOT(x, install("pi"));
      |         ^
solve_subset.c:108:9: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  108 |     tmp = GET_SLOT(x, install("px"));
      |         ^
solve_subset.c:110:9: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  110 |     tmp = GET_SLOT(x, install("s"));
      |         ^
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from solve_subset.c:53:
solve_subset.c: In function ‘tmb_CHMfactor_solve’:
solve_subset.c:146:23: warning: implicit declaration of function ‘NEW_NUMERIC’ [-Wimplicit-function-declaration]
  146 |     SEXP x = PROTECT( NEW_NUMERIC( n ) );
      |                       ^~~~~~~~~~~
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
solve_subset.c:146:23: warning: passing argument 1 of ‘Rf_protect’ makes pointer from integer without a cast [-Wint-conversion]
  146 |     SEXP x = PROTECT( NEW_NUMERIC( n ) );
      |                       ^~~~~~~~~~~~~~~~
      |                       |
      |                       int
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from solve_subset.c:53:
/usr/local/lib/R/include/Rinternals.h:1101:17: note: expected ‘SEXP’ {aka ‘struct SEXPREC *’} but argument is of type ‘int’
 1101 | SEXP Rf_protect(SEXP);
      |                 ^~~~
In file included from /usr/local/lib/R/site-library/Matrix/include/Matrix.h:4,
                 from solve_subset.c:53:
solve_subset.c: In function ‘match_pattern’:
solve_subset.c:334:15: warning: implicit declaration of function ‘NEW_INTEGER’; did you mean ‘NA_INTEGER’? [-Wimplicit-function-declaration]
  334 |   PROTECT(ans=NEW_INTEGER(A->nzmax));
      |               ^~~~~~~~~~~
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
solve_subset.c:334:14: warning: assignment to ‘SEXP’ {aka ‘struct SEXPREC *’} from ‘int’ makes pointer from integer without a cast [-Wint-conversion]
  334 |   PROTECT(ans=NEW_INTEGER(A->nzmax));
      |              ^
/usr/local/lib/R/include/Rinternals.h:368:36: note: in definition of macro ‘PROTECT’
  368 | #define PROTECT(s)      Rf_protect(s)
      |                                    ^
gcc -I"/usr/local/lib/R/include" -DNDEBUG  -I'/usr/local/lib/R/site-library/Matrix/include' -I/usr/local/include   -fopenmp -fpic  -g -O2  -c utils.c -o utils.o
gcc -shared -L/usr/local/lib/R/lib -L/usr/local/lib -o TMB.so external_metis.o init.o local_stubs.o solve_subset.o utils.o -llapack -L/usr/local/lib/R/lib -lRblas -lgfortran -lm -lquadmath -fopenmp -L/usr/local/lib/R/lib -lR
make[1]: Leaving directory '/tmp/RtmpNffoEw/R.INSTALL137d4489d8994/TMB/src'
installing to /usr/local/lib/R/site-library/00LOCK-TMB/00new/TMB/libs
** R
** inst
** byte-compile and prepare package for lazy loading
Loading required package: utils
Tracing function "install.packages" in package "utils"
** help
*** installing help indices
** building package indices
Loading required package: utils
Tracing function "install.packages" in package "utils"
** testing if installed package can be loaded from temporary location
Loading required package: utils
Tracing function "install.packages" in package "utils"
Error: package or namespace load failed for ‘TMB’:
 .onLoad failed in loadNamespace() for 'TMB', details:
  call: dyn.load(file, DLLpath = DLLpath, ...)
  error: unable to load shared object '/usr/local/lib/R/site-library/00LOCK-TMB/00new/TMB/libs/TMB.so':
  /usr/local/lib/R/site-library/00LOCK-TMB/00new/TMB/libs/TMB.so: undefined symbol: NEW_INTEGER
Error: loading failed
Execution halted
ERROR: loading failed
* removing ‘/usr/local/lib/R/site-library/TMB’
* restoring previous ‘/usr/local/lib/R/site-library/TMB’
make: *** [Makefile:24: install] Error 1
kaskr commented 1 year ago

Thanks @bbolker - should be fixed now. It would be nice if you could confirm that it works...

bbolker commented 1 year ago

Yes, thank you.