Hello folks,
I'm embarking on a project to embed R into the Apache web server, and I'd like your help. Currently, I'm looking for a way for R code to call back into a shared library from which the R shared library was loaded.
Essentially, apache starts and loads mod_R.so which runs an initialization routine which calls Rf_initEmbeddedR() and the following code:
/* override to call apache library write routine */ ptr_R_WriteConsole = Rapache_WriteConsole;
/* load source(). I assume this is appropriate. I could
always move this to the code that handles the requst */
PROTECT(R_source_fun = allocVector(LANGSXP, 2));
SETCAR(R_source_fun, Rf_install("source"));
SETCAR(CDR(R_source_fun), R_source_arg = NEW_CHARACTER(1));Then each request is serviced by the following code:
/* r is apache request, r->filename is url translated to
apsolute path to filename */
SET_STRING_ELT(R_source_arg, 0, COPY_TO_USER_STRING(r->filename));
errorOccurred=1;
/* R code: source(filename) */
val = R_tryEval(R_source_fun,NULL,&errorOccurred);
if (errorOccurred){
// Send a message to stderr (apache redirects this to the error log)
fprintf(stderr,"apache2_mod_R: source(%s) failed!\n",r->filename);
fflush(stderr);
} else {
return ok;
}I would like to write R routines which hook into apache specific library calls to get at the CGI and system data. The apache specific calls need a request structure (the r variable) which is available in mod_R.so (already loaded) on each request. If I use the .C, .Call, or .External interface, then I must load a shared library with the appropriate routines. I would like to place those routines in mod_R.so to take advantage of access to the request structure. Is this doable?
Another approach is to turn the CGI data into R variables and just place those in the top level environment, foregoing any callbacks into apache again until the R code has run to completion. This is similar to how PHP is embedded into Apache. Other languages that _do_ provide callbacks into apache include Perl, Python, and Ruby.
-- Jeffrey Horner Computer Systems Analyst School of Medicine 615-322-8606 Department of Biostatistics Vanderbilt University
______________________________________________ [EMAIL PROTECTED] mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
