[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[netCDF #PUM-976132]: Compile problems with NetCDF 3.6.3 and gcc's
- Subject: [netCDF #PUM-976132]: Compile problems with NetCDF 3.6.3 and gcc's
- Date: Thu, 24 Jul 2008 22:31:03 -0600
Sean,
With the updated genlib.c, "make check" now works fine on Mac OS X 10.5.4
using gcc 4.2.1 with -D_FORTIFY_SOURCE=2. The problem you found was the
only thing caught by the FORTIFY_SOURCE checks. We're currently having a
little trouble generating daily snapshots, so I've attached the updated
genlib.c file, intended to replace ncgen/genlib.c in either the netcdf-3.6.3
or netcdf-4.0 release.
--Russ
Russ Rew UCAR Unidata Program
address@hidden http://www.unidata.ucar.edu
Ticket Details
===================
Ticket ID: PUM-976132
Department: Support netCDF
Priority: Normal
Status: Closed
/*********************************************************************
* Copyright 1993, UCAR/Unidata
* See netcdf/COPYRIGHT file for copying and redistribution conditions.
* $Header: /upc/share/CVS/netcdf-3/ncgen/genlib.c,v 1.52 2008/07/24 23:29:09
russ Exp $
*********************************************************************/
#include <config.h>
#include <stdio.h>
#include <stdlib.h>
#include <assert.h>
#include <string.h>
#include <ctype.h> /* for isprint() */
#ifndef NO_STDARG
#include <stdarg.h>
#else
/* try varargs instead */
#include <varargs.h>
#endif /* !NO_STDARG */
#include <netcdf.h>
#include "generic.h"
#include "ncgen.h"
#include "genlib.h"
extern char *netcdf_name; /* output netCDF filename, if on command line. */
extern int netcdf_flag;
extern int c_flag;
extern int fortran_flag;
extern int cmode_modifier;
extern int nofill_flag;
int lineno = 1;
int derror_count = 0;
/* create netCDF from in-memory structure */
static void
gen_netcdf(
char *filename) /* name for output netcdf file */
{
int idim, ivar, iatt;
int dimid;
int varid;
int stat;
stat = nc_create(filename, cmode_modifier, &ncid);
check_err(stat);
/* define dimensions from info in dims array */
for (idim = 0; idim < ndims; idim++) {
stat = nc_def_dim(ncid, dims[idim].name, dims[idim].size, &dimid);
check_err(stat);
}
/* define variables from info in vars array */
for (ivar = 0; ivar < nvars; ivar++) {
stat = nc_def_var(ncid,
vars[ivar].name,
vars[ivar].type,
vars[ivar].ndims,
vars[ivar].dims,
&varid);
check_err(stat);
}
/* define attributes from info in atts array */
for (iatt = 0; iatt < natts; iatt++) {
varid = (atts[iatt].var == -1) ? NC_GLOBAL : atts[iatt].var;
switch(atts[iatt].type) {
case NC_BYTE:
stat = nc_put_att_schar(ncid, varid, atts[iatt].name,
atts[iatt].type, atts[iatt].len,
(signed char *) atts[iatt].val);
break;
case NC_CHAR:
stat = nc_put_att_text(ncid, varid, atts[iatt].name,
atts[iatt].len,
(char *) atts[iatt].val);
break;
case NC_SHORT:
stat = nc_put_att_short(ncid, varid, atts[iatt].name,
atts[iatt].type, atts[iatt].len,
(short *) atts[iatt].val);
break;
case NC_INT:
stat = nc_put_att_int(ncid, varid, atts[iatt].name,
atts[iatt].type, atts[iatt].len,
(int *) atts[iatt].val);
break;
case NC_FLOAT:
stat = nc_put_att_float(ncid, varid, atts[iatt].name,
atts[iatt].type, atts[iatt].len,
(float *) atts[iatt].val);
break;
case NC_DOUBLE:
stat = nc_put_att_double(ncid, varid, atts[iatt].name,
atts[iatt].type, atts[iatt].len,
(double *) atts[iatt].val);
break;
default:
stat = NC_EBADTYPE;
}
check_err(stat);
}
if (nofill_flag) {
stat = nc_set_fill(ncid, NC_NOFILL, 0); /* don't initialize with fill
values */
check_err(stat);
}
stat = nc_enddef(ncid);
check_err(stat);
}
/*
* Given a netcdf type, a pointer to a vector of values of that type,
* and the index of the vector element desired, returns a pointer to a
* malloced string representing the value in C.
*/
static char *
cstring(
nc_type type, /* netCDF type code */
void *valp, /* pointer to vector of values */
int num) /* element of vector desired */
{
static char *cp, *sp, ch;
signed char *bytep;
short *shortp;
int *intp;
float *floatp;
double *doublep;
switch (type) {
case NC_CHAR:
sp = cp = (char *) emalloc (7);
*cp++ = '\'';
ch = *((char *)valp + num);
switch (ch) {
case '\b': *cp++ = '\\'; *cp++ = 'b'; break;
case '\f': *cp++ = '\\'; *cp++ = 'f'; break;
case '\n': *cp++ = '\\'; *cp++ = 'n'; break;
case '\r': *cp++ = '\\'; *cp++ = 'r'; break;
case '\t': *cp++ = '\\'; *cp++ = 't'; break;
case '\v': *cp++ = '\\'; *cp++ = 'v'; break;
case '\\': *cp++ = '\\'; *cp++ = '\\'; break;
case '\'': *cp++ = '\\'; *cp++ = '\''; break;
default:
if (!isprint((unsigned char)ch)) {
static char octs[] = "01234567";
int rem = ((unsigned char)ch)%64;
*cp++ = '\\';
*cp++ = octs[((unsigned char)ch)/64]; /* to get, e.g. '\177' */
*cp++ = octs[rem/8];
*cp++ = octs[rem%8];
} else {
*cp++ = ch;
}
break;
}
*cp++ = '\'';
*cp = '\0';
return sp;
case NC_BYTE:
cp = (char *) emalloc (7);
bytep = (signed char *)valp;
/* Need to convert '\377' to -1, for example, on all platforms */
(void) sprintf(cp,"%d", (signed char) *(bytep+num));
return cp;
case NC_SHORT:
cp = (char *) emalloc (10);
shortp = (short *)valp;
(void) sprintf(cp,"%d",* (shortp + num));
return cp;
case NC_INT:
cp = (char *) emalloc (20);
intp = (int *)valp;
(void) sprintf(cp,"%d",* (intp + num));
return cp;
case NC_FLOAT:
cp = (char *) emalloc (20);
floatp = (float *)valp;
(void) sprintf(cp,"%.8g",* (floatp + num));
return cp;
case NC_DOUBLE:
cp = (char *) emalloc (20);
doublep = (double *)valp;
(void) sprintf(cp,"%.16g",* (doublep + num));
return cp;
default:
derror("cstring: bad type code");
return 0;
}
}
/*
* Generate C code for creating netCDF from in-memory structure.
*/
static void
gen_c(
const char *filename)
{
int idim, ivar, iatt, jatt, maxdims;
int vector_atts;
char *val_string;
char stmnt[C_MAX_STMNT];
/* wrap in main program */
cline("#include <stdio.h>");
cline("#include <stdlib.h>");
cline("#include <netcdf.h>");
cline("");
cline("void");
cline("check_err(const int stat, const int line, const char *file) {");
cline(" if (stat != NC_NOERR) {");
cline(" (void) fprintf(stderr, \"line %d of %s: %s\\n\", line, file,
nc_strerror(stat));");
cline(" exit(1);");
cline(" }");
cline("}");
cline("");
cline("int");
sprintf(stmnt, "main() {\t\t\t/* create %s */", filename);
cline(stmnt);
/* create necessary declarations */
cline("");
cline(" int stat;\t\t\t/* return status */");
cline(" int ncid;\t\t\t/* netCDF id */");
if (ndims > 0) {
cline("");
cline(" /* dimension ids */");
for (idim = 0; idim < ndims; idim++) {
sprintf(stmnt, " int %s_dim;", dims[idim].lname);
cline(stmnt);
}
cline("");
cline(" /* dimension lengths */");
for (idim = 0; idim < ndims; idim++) {
if (dims[idim].size == NC_UNLIMITED) {
sprintf(stmnt, " size_t %s_len = NC_UNLIMITED;",
dims[idim].lname);
} else {
sprintf(stmnt, " size_t %s_len = %lu;",
dims[idim].lname,
(unsigned long) dims[idim].size);
}
cline(stmnt);
}
}
maxdims = 0; /* most dimensions of any variable */
for (ivar = 0; ivar < nvars; ivar++)
if (vars[ivar].ndims > maxdims)
maxdims = vars[ivar].ndims;
if (nvars > 0) {
cline("");
cline(" /* variable ids */");
for (ivar = 0; ivar < nvars; ivar++) {
sprintf(stmnt, " int %s_id;", vars[ivar].lname);
cline(stmnt);
}
cline("");
cline(" /* rank (number of dimensions) for each variable */");
for (ivar = 0; ivar < nvars; ivar++) {
sprintf(stmnt, "# define RANK_%s %d", vars[ivar].lname,
vars[ivar].ndims);
cline(stmnt);
}
if (maxdims > 0) { /* we have dimensioned variables */
cline("");
cline(" /* variable shapes */");
for (ivar = 0; ivar < nvars; ivar++) {
if (vars[ivar].ndims > 0) {
sprintf(stmnt, " int %s_dims[RANK_%s];",
vars[ivar].lname, vars[ivar].lname);
cline(stmnt);
}
}
}
}
/* determine if we need any attribute vectors */
vector_atts = 0;
for (iatt = 0; iatt < natts; iatt++) {
if (atts[iatt].type != NC_CHAR) {
vector_atts = 1;
break;
}
}
if (vector_atts) {
cline("");
cline(" /* attribute vectors */");
for (iatt = 0; iatt < natts; iatt++) {
if (atts[iatt].type != NC_CHAR) {
sprintf(stmnt,
" %s %s_%s[%lu];",
ncatype(atts[iatt].type),
atts[iatt].var == -1 ? "cdf" : vars[atts[iatt].var].lname,
atts[iatt].lname,
(unsigned long) atts[iatt].len);
cline(stmnt);
}
}
}
/* create netCDF file, uses NC_CLOBBER mode */
cline("");
cline(" /* enter define mode */");
if (!cmode_modifier) {
sprintf(stmnt,
" stat = nc_create(\"%s\", NC_CLOBBER, &ncid);",
filename);
} else if (cmode_modifier & NC_64BIT_OFFSET) {
sprintf(stmnt,
" stat = nc_create(\"%s\", NC_CLOBBER|NC_64BIT_OFFSET,
&ncid);",
filename);
#ifdef USE_NETCDF4
} else if (cmode_modifier & NC_CLASSIC_MODEL) {
sprintf(stmnt,
" stat = nc_create(\"%s\",
NC_CLOBBER|NC_NETCDF4|NC_CLASSIC_MODEL, &ncid);",
filename);
} else if (cmode_modifier & NC_NETCDF4) {
sprintf(stmnt,
" stat = nc_create(\"%s\", NC_CLOBBER|NC_NETCDF4, &ncid);",
filename);
#endif
} else {
derror("unknown cmode modifier");
}
cline(stmnt);
cline(" check_err(stat,__LINE__,__FILE__);");
/* define dimensions from info in dims array */
if (ndims > 0) {
cline("");
cline(" /* define dimensions */");
}
for (idim = 0; idim < ndims; idim++) {
sprintf(stmnt,
" stat = nc_def_dim(ncid, \"%s\", %s_len, &%s_dim);",
dims[idim].name, dims[idim].lname, dims[idim].lname);
cline(stmnt);
cline(" check_err(stat,__LINE__,__FILE__);");
}
/* define variables from info in vars array */
if (nvars > 0) {
cline("");
cline(" /* define variables */");
for (ivar = 0; ivar < nvars; ivar++) {
cline("");
for (idim = 0; idim < vars[ivar].ndims; idim++) {
sprintf(stmnt,
" %s_dims[%d] = %s_dim;",
vars[ivar].lname,
idim,
dims[vars[ivar].dims[idim]].lname);
cline(stmnt);
}
if (vars[ivar].ndims > 0) { /* a dimensioned variable */
sprintf(stmnt,
" stat = nc_def_var(ncid, \"%s\", %s, RANK_%s,
%s_dims, &%s_id);",
vars[ivar].name,
nctype(vars[ivar].type),
vars[ivar].lname,
vars[ivar].lname,
vars[ivar].lname);
} else { /* a scalar */
sprintf(stmnt,
" stat = nc_def_var(ncid, \"%s\", %s, RANK_%s, 0,
&%s_id);",
vars[ivar].name,
nctype(vars[ivar].type),
vars[ivar].lname,
vars[ivar].lname);
}
cline(stmnt);
cline(" check_err(stat,__LINE__,__FILE__);");
}
}
/* define attributes from info in atts array */
if (natts > 0) {
cline("");
cline(" /* assign attributes */");
for (iatt = 0; iatt < natts; iatt++) {
if (atts[iatt].type == NC_CHAR) { /* string */
val_string = cstrstr((char *) atts[iatt].val, atts[iatt].len);
sprintf(stmnt,
" stat = nc_put_att_text(ncid, %s%s, \"%s\", %lu,
%s);",
atts[iatt].var == -1 ? "NC_GLOBAL" :
vars[atts[iatt].var].lname,
atts[iatt].var == -1 ? "" : "_id",
atts[iatt].name,
(unsigned long) atts[iatt].len,
val_string);
cline(stmnt);
free (val_string);
}
else { /* vector attribute */
for (jatt = 0; jatt < atts[iatt].len ; jatt++) {
val_string = cstring(atts[iatt].type,atts[iatt].val,jatt);
sprintf(stmnt, " %s_%s[%d] = %s;",
atts[iatt].var == -1 ? "cdf" :
vars[atts[iatt].var].lname,
atts[iatt].lname,
jatt,
val_string);
cline(stmnt);
free (val_string);
}
sprintf(stmnt,
" stat = nc_put_att_%s(ncid, %s%s, \"%s\", %s, %lu,
%s_%s);",
ncatype(atts[iatt].type),
atts[iatt].var == -1 ? "NC_GLOBAL" :
vars[atts[iatt].var].lname,
atts[iatt].var == -1 ? "" : "_id",
atts[iatt].name,
nctype(atts[iatt].type),
(unsigned long) atts[iatt].len,
atts[iatt].var == -1 ? "cdf" :
vars[atts[iatt].var].lname,
atts[iatt].lname);
cline(stmnt);
}
cline(" check_err(stat,__LINE__,__FILE__);");
}
}
if (nofill_flag) {
cline(" /* don't initialize variables with fill values */");
cline(" stat = nc_set_fill(ncid, NC_NOFILL, 0);");
cline(" check_err(stat,__LINE__,__FILE__);");
}
cline("");
cline(" /* leave define mode */");
cline(" stat = nc_enddef (ncid);");
cline(" check_err(stat,__LINE__,__FILE__);");
}
/* return Fortran type name for netCDF type, given type code */
static const char *
ncftype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "integer";
case NC_CHAR:
return "character";
case NC_SHORT:
return "integer";
case NC_INT:
#ifdef MSDOS
return "integer*4";
#else
return "integer";
#endif
case NC_FLOAT:
return "real";
#if defined(_CRAY) && !defined(__crayx1)
case NC_DOUBLE:
return "real"; /* we don't support CRAY 128-bit doubles */
#else
case NC_DOUBLE:
return "double precision";
#endif
default:
derror("ncftype: bad type code");
return 0;
}
}
/* return Fortran type suffix for netCDF type, given type code */
const char *
nfstype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "int1";
case NC_CHAR:
return "text";
case NC_SHORT:
return "int2";
case NC_INT:
return "int";
case NC_FLOAT:
return "real";
case NC_DOUBLE:
return "double";
default:
derror("nfstype: bad type code");
return 0;
}
}
/* Return Fortran function suffix for netCDF type, given type code.
* This should correspond to the Fortran type name in ncftype().
*/
const char *
nfftype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "int";
case NC_CHAR:
return "text";
case NC_SHORT:
return "int";
case NC_INT:
return "int";
case NC_FLOAT:
return "real";
#if defined(_CRAY) && !defined(__crayx1)
case NC_DOUBLE:
return "real"; /* we don't support CRAY 128-bit doubles */
#else
case NC_DOUBLE:
return "double";
#endif
default:
derror("nfstype: bad type code");
return 0;
}
}
/* return FORTRAN name for netCDF type, given type code */
static const char *
ftypename(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "NF_INT1";
case NC_CHAR:
return "NF_CHAR";
case NC_SHORT:
return "NF_INT2";
case NC_INT:
return "NF_INT";
case NC_FLOAT:
return "NF_REAL";
case NC_DOUBLE:
return "NF_DOUBLE";
default:
derror("ftypename: bad type code");
return 0;
}
}
/*
* Generate FORTRAN code for creating netCDF from in-memory structure.
*/
static void
gen_fortran(
const char *filename)
{
int idim, ivar, iatt, jatt, itype, maxdims;
int vector_atts;
char *val_string;
char stmnt[FORT_MAX_STMNT];
char s2[NC_MAX_NAME + 10];
char *sp;
/* Need how many netCDF types there are, because we create an array
* for each type of attribute. */
int ntypes = 6; /* number of netCDF types, NC_BYTE, ... */
nc_type types[6]; /* at least ntypes */
size_t max_atts[NC_DOUBLE + 1];
types[0] = NC_BYTE;
types[1] = NC_CHAR;
types[2] = NC_SHORT;
types[3] = NC_INT;
types[4] = NC_FLOAT;
types[5] = NC_DOUBLE;
fline("program fgennc");
fline("include 'netcdf.inc'");
/* create necessary declarations */
fline("* error status return");
fline("integer iret");
fline("* netCDF id");
fline("integer ncid");
if (nofill_flag) {
fline("* to save old fill mode before changing it temporarily");
fline("integer oldmode");
}
if (ndims > 0) {
fline("* dimension ids");
for (idim = 0; idim < ndims; idim++) {
sprintf(stmnt, "integer %s_dim", dims[idim].lname);
fline(stmnt);
}
fline("* dimension lengths");
for (idim = 0; idim < ndims; idim++) {
sprintf(stmnt, "integer %s_len", dims[idim].lname);
fline(stmnt);
}
for (idim = 0; idim < ndims; idim++) {
if (dims[idim].size == NC_UNLIMITED) {
sprintf(stmnt, "parameter (%s_len = NF_UNLIMITED)",
dims[idim].lname);
} else {
sprintf(stmnt, "parameter (%s_len = %lu)",
dims[idim].lname,
(unsigned long) dims[idim].size);
}
fline(stmnt);
}
}
maxdims = 0; /* most dimensions of any variable */
for (ivar = 0; ivar < nvars; ivar++)
if (vars[ivar].ndims > maxdims)
maxdims = vars[ivar].ndims;
if (nvars > 0) {
fline("* variable ids");
for (ivar = 0; ivar < nvars; ivar++) {
sprintf(stmnt, "integer %s_id", vars[ivar].lname);
fline(stmnt);
}
fline("* rank (number of dimensions) for each variable");
for (ivar = 0; ivar < nvars; ivar++) {
sprintf(stmnt, "integer %s_rank", vars[ivar].lname);
fline(stmnt);
}
for (ivar = 0; ivar < nvars; ivar++) {
sprintf(stmnt, "parameter (%s_rank = %d)", vars[ivar].lname,
vars[ivar].ndims);
fline(stmnt);
}
fline("* variable shapes");
for (ivar = 0; ivar < nvars; ivar++) {
if (vars[ivar].ndims > 0) {
sprintf(stmnt, "integer %s_dims(%s_rank)",
vars[ivar].lname, vars[ivar].lname);
fline(stmnt);
}
}
}
/* declarations for variables to be initialized */
if (nvars > 0) { /* we have variables */
fline("* data variables");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
/* Generate declarations here for non-record data variables only.
Record variables are declared in separate subroutine later,
when we know how big they are. */
if (v->ndims > 0 && v->dims[0] == rec_dim) {
continue;
}
/* Make declarations for non-text variables only;
for text variables, just include string in nf_put_var call */
if (v->type == NC_CHAR) {
continue;
}
if (v->ndims == 0) { /* scalar */
sprintf(stmnt, "%s %s", ncftype(v->type),
v->lname);
} else {
sprintf(stmnt, "%s %s(", ncftype(v->type),
v->lname);
/* reverse dimensions for FORTRAN */
for (idim = v->ndims-1; idim >= 0; idim--) {
sprintf(s2, "%s_len, ",
dims[v->dims[idim]].lname);
strcat(stmnt, s2);
}
sp = strrchr(stmnt, ',');
if(sp != NULL) {
*sp = '\0';
}
strcat(stmnt, ")");
}
fline(stmnt);
}
}
/* determine what attribute vectors needed */
for (itype = 0; itype < ntypes; itype++)
max_atts[(int)types[itype]] = 0;
vector_atts = 0;
for (iatt = 0; iatt < natts; iatt++) {
if (atts[iatt].len > max_atts[(int) atts[iatt].type]) {
max_atts[(int)atts[iatt].type] = atts[iatt].len;
vector_atts = 1;
}
}
if (vector_atts) {
fline("* attribute vectors");
for (itype = 0; itype < ntypes; itype++) {
if (types[itype] != NC_CHAR && max_atts[(int)types[itype]] > 0) {
sprintf(stmnt, "%s %sval(%lu)", ncftype(types[itype]),
nfstype(types[itype]),
(unsigned long) max_atts[(int)types[itype]]);
fline(stmnt);
}
}
}
/* create netCDF file, uses NC_CLOBBER mode */
fline("* enter define mode");
if (!cmode_modifier) {
sprintf(stmnt, "iret = nf_create(\'%s\', NF_CLOBBER, ncid)", filename);
} else if (cmode_modifier & NC_64BIT_OFFSET) {
sprintf(stmnt, "iret = nf_create(\'%s\',
OR(NF_CLOBBER,NF_64BIT_OFFSET), ncid)", filename);
#ifdef USE_NETCDF4
} else if (cmode_modifier & NC_CLASSIC_MODEL) {
sprintf(stmnt, "iret = nf_create(\'%s\',
OR(NF_CLOBBER,NC_NETCDF4,NC_CLASSIC_MODEL), ncid)", filename);
} else if (cmode_modifier & NC_NETCDF4) {
sprintf(stmnt, "iret = nf_create(\'%s\', OR(NF_CLOBBER,NF_NETCDF4),
ncid)", filename);
#endif
} else {
derror("unknown cmode modifier");
}
fline(stmnt);
fline("call check_err(iret)");
/* define dimensions from info in dims array */
if (ndims > 0)
fline("* define dimensions");
for (idim = 0; idim < ndims; idim++) {
if (dims[idim].size == NC_UNLIMITED)
sprintf(stmnt, "iret = nf_def_dim(ncid, \'%s\', NF_UNLIMITED,
%s_dim)",
dims[idim].name, dims[idim].lname);
else
sprintf(stmnt, "iret = nf_def_dim(ncid, \'%s\', %lu, %s_dim)",
dims[idim].name, (unsigned long) dims[idim].size,
dims[idim].lname);
fline(stmnt);
fline("call check_err(iret)");
}
/* define variables from info in vars array */
if (nvars > 0) {
fline("* define variables");
for (ivar = 0; ivar < nvars; ivar++) {
for (idim = 0; idim < vars[ivar].ndims; idim++) {
sprintf(stmnt, "%s_dims(%d) = %s_dim",
vars[ivar].lname,
vars[ivar].ndims - idim, /* reverse dimensions */
dims[vars[ivar].dims[idim]].lname);
fline(stmnt);
}
if (vars[ivar].ndims > 0) { /* a dimensioned variable */
sprintf(stmnt,
"iret = nf_def_var(ncid, \'%s\', %s, %s_rank, %s_dims,
%s_id)",
vars[ivar].name,
ftypename(vars[ivar].type),
vars[ivar].lname,
vars[ivar].lname,
vars[ivar].lname);
} else { /* a scalar */
sprintf(stmnt,
"iret = nf_def_var(ncid, \'%s\', %s, %s_rank, 0,
%s_id)",
vars[ivar].name,
ftypename(vars[ivar].type),
vars[ivar].lname,
vars[ivar].lname);
}
fline(stmnt);
fline("call check_err(iret)");
}
}
/* define attributes from info in atts array */
if (natts > 0) {
fline("* assign attributes");
for (iatt = 0; iatt < natts; iatt++) {
if (atts[iatt].type == NC_CHAR) { /* string */
val_string = fstrstr((char *) atts[iatt].val, atts[iatt].len);
sprintf(stmnt,
"iret = nf_put_att_text(ncid, %s%s, \'%s\', %lu, %s)",
atts[iatt].var == -1 ? "NF_GLOBAL" :
vars[atts[iatt].var].lname,
atts[iatt].var == -1 ? "" : "_id",
atts[iatt].name,
(unsigned long) atts[iatt].len,
val_string);
fline(stmnt);
fline("call check_err(iret)");
free(val_string);
} else {
for (jatt = 0; jatt < atts[iatt].len ; jatt++) {
val_string = fstring(atts[iatt].type,atts[iatt].val,jatt);
sprintf(stmnt, "%sval(%d) = %s",
nfstype(atts[iatt].type),
jatt+1,
val_string);
fline(stmnt);
free (val_string);
}
sprintf(stmnt,
"iret = nf_put_att_%s(ncid, %s%s, \'%s\', %s, %lu,
%sval)",
nfftype(atts[iatt].type),
atts[iatt].var == -1 ? "NCGLOBAL" :
vars[atts[iatt].var].lname,
atts[iatt].var == -1 ? "" : "_id",
atts[iatt].name,
ftypename(atts[iatt].type),
(unsigned long) atts[iatt].len,
nfstype(atts[iatt].type));
fline(stmnt);
fline("call check_err(iret)");
}
}
}
if (nofill_flag) {
fline("* don't initialize variables with fill values");
fline("iret = nf_set_fill(ncid, NF_NOFILL, oldmode)");
fline("call check_err(iret)");
}
fline("* leave define mode");
fline("iret = nf_enddef(ncid)");
fline("call check_err(iret)");
}
/*
* Output a C statement.
*/
void
cline(
const char *stmnt)
{
FILE *cout = stdout;
fputs(stmnt, cout);
fputs("\n", cout);
}
/*
* From a long line FORTRAN statment, generates the necessary FORTRAN
* lines with continuation characters in column 6. If stmnt starts with "*",
* it is treated as a one-line comment. Statement labels are *not* handled,
* but since we don't generate any labels, we don't care.
*/
void
fline(
const char *stmnt)
{
FILE *fout = stdout;
int len = (int) strlen(stmnt);
int line = 0;
static char cont[] = { /* continuation characters */
' ', '1', '2', '3', '4', '5', '6', '7', '8', '9',
'+', '1', '2', '3', '4', '5', '6', '7', '8', '9',
'+', '1', '2', '3', '4', '5', '6', '7', '8', '9'};
if(stmnt[0] == '*') {
fputs(stmnt, fout);
fputs("\n", fout);
return;
}
while (len > 0) {
if (line >= FORT_MAX_LINES)
derror("FORTRAN statement too long: %s",stmnt);
(void) fprintf(fout, " %c", cont[line++]);
(void) fprintf(fout, "%.66s\n", stmnt);
len -= 66;
if (len > 0)
stmnt += 66;
}
}
/* return C name for netCDF type, given type code */
const char *
nctype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "NC_BYTE";
case NC_CHAR:
return "NC_CHAR";
case NC_SHORT:
return "NC_SHORT";
case NC_INT:
return "NC_INT";
case NC_FLOAT:
return "NC_FLOAT";
case NC_DOUBLE:
return "NC_DOUBLE";
default:
derror("nctype: bad type code");
return 0;
}
}
/*
* Return C type name for netCDF type, given type code.
*/
const char *
ncctype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "signed char";
case NC_CHAR:
return "char";
case NC_SHORT:
return "short";
case NC_INT:
return "int";
case NC_FLOAT:
return "float";
case NC_DOUBLE:
return "double";
default:
derror("ncctype: bad type code");
return 0;
}
}
/*
* Return C type name for netCDF type suffix, given type code.
*/
const char *
ncstype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "schar";
case NC_CHAR:
return "text";
case NC_SHORT:
return "short";
case NC_INT:
return "int";
case NC_FLOAT:
return "float";
case NC_DOUBLE:
return "double";
default:
derror("ncstype: bad type code");
return 0;
}
}
/*
* Return C type name for netCDF attribute container type, given type code.
*/
const char *
ncatype(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return "int"; /* avoids choosing between uchar and schar */
case NC_CHAR:
return "text";
case NC_SHORT:
return "short";
case NC_INT:
return "int";
case NC_FLOAT:
return "float";
case NC_DOUBLE:
return "double";
default:
derror("ncatype: bad type code");
return 0;
}
}
/* return internal size for values of specified netCDF type */
size_t
nctypesize(
nc_type type) /* netCDF type code */
{
switch (type) {
case NC_BYTE:
return sizeof(char);
case NC_CHAR:
return sizeof(char);
case NC_SHORT:
return sizeof(short);
case NC_INT:
return sizeof(int);
case NC_FLOAT:
return sizeof(float);
case NC_DOUBLE:
return sizeof(double);
default:
derror("nctypesize: bad type code");
return 0;
}
}
/*
* Given a netcdf numeric type, a pointer to a vector of values of that
* type, and the index of the vector element desired, returns a pointer
* to a malloced string representing the value in FORTRAN. Since this
* may be used in a DATA statement, it must not include non-constant
* expressions, such as "char(26)".
*/
char *
fstring(
nc_type type, /* netCDF type code */
void *valp, /* pointer to vector of values */
int num) /* element of vector desired */
{
static char *cp;
signed char *schp;
short *shortp;
int *intp;
float *floatp;
double *doublep;
switch (type) {
case NC_BYTE:
cp = (char *) emalloc (10);
schp = (signed char *)valp;
sprintf(cp,"%d", schp[num]);
return cp;
case NC_SHORT:
cp = (char *) emalloc (10);
shortp = (short *)valp;
(void) sprintf(cp,"%d",* (shortp + num));
return cp;
case NC_INT:
cp = (char *) emalloc (20);
intp = (int *)valp;
(void) sprintf(cp,"%d",* (intp + num));
return cp;
case NC_FLOAT:
cp = (char *) emalloc (20);
floatp = (float *)valp;
(void) sprintf(cp,"%.8g",* (floatp + num));
return cp;
case NC_DOUBLE:
cp = (char *) emalloc (25);
doublep = (double *)valp;
(void) sprintf(cp,"%.16g",* (doublep + num));
expe2d(cp); /* change 'e' to 'd' in exponent */
return cp;
default:
derror("fstring: bad type code");
return 0;
}
}
/*
* Given a pointer to a counted string, returns a pointer to a malloced string
* representing the string as a C constant.
*/
char *
cstrstr(
const char *valp, /* pointer to vector of characters*/
size_t len) /* number of characters in valp */
{
static char *sp;
char *cp;
char *istr, *istr0; /* for null-terminated copy */
int ii;
if(4*len+3 != (unsigned)(4*len+3)) {
derror("too much character data!");
exit(9);
}
sp = cp = (char *) emalloc(4*len+3);
if(len == 1 && *valp == 0) { /* empty string */
strcpy(sp,"\"\"");
return sp;
}
istr0 = istr = (char *) emalloc(len + 1);
for(ii = 0; ii < len; ii++) {
istr[ii] = valp[ii];
}
istr[len] = '\0';
*cp++ = '"';
for(ii = 0; ii < len; ii++) {
switch (*istr) {
case '\0': *cp++ = '\\'; *cp++ = '0'; *cp++ = '0'; *cp++ = '0'; break;
case '\b': *cp++ = '\\'; *cp++ = 'b'; break;
case '\f': *cp++ = '\\'; *cp++ = 'f'; break;
case '\n': *cp++ = '\\'; *cp++ = 'n'; break;
case '\r': *cp++ = '\\'; *cp++ = 'r'; break;
case '\t': *cp++ = '\\'; *cp++ = 't'; break;
case '\v': *cp++ = '\\'; *cp++ = 'v'; break;
case '\\': *cp++ = '\\'; *cp++ = '\\'; break;
case '\"': *cp++ = '\\'; *cp++ = '\"'; break;
default:
if (!isprint((unsigned char)*istr)) {
static char octs[] = "01234567";
int rem = ((unsigned char)*istr)%64;
*cp++ = '\\';
*cp++ = octs[((unsigned char)*istr)/64]; /* to get, e.g. '\177'
*/
*cp++ = octs[rem/8];
*cp++ = octs[rem%8];
} else {
*cp++ = *istr;
}
break;
}
istr++;
}
*cp++ = '"';
*cp = '\0';
free(istr0);
return sp;
}
/* Given a pointer to a counted string (not necessarily
* null-terminated), returns a pointer to a malloced string representing
* the string as a FORTRAN string expression. For example, the string
* "don't" would yield the FORTRAN string "'don''t'", and the string
* "ab\ncd" would yield "'ab'//char(10)//'cd'". The common
* interpretation of "\"-escaped characters is non-standard, so the
* generated Fortran may require adjustment in compilers that don't
* recognize "\" as anything special in a character context. */
char *
fstrstr(
const char *str, /* pointer to vector of characters */
size_t ilen) /* number of characters in istr */
{
static char *ostr;
char *cp, tstr[12];
int was_print = 0; /* true if last character was printable */
char *istr, *istr0; /* for null-terminated copy */
int ii;
if(12*ilen != (size_t)(12*ilen)) {
derror("too much character data!");
exit(9);
}
istr0 = istr = (char *) emalloc(ilen + 1);
for(ii = 0; ii < ilen; ii++) {
istr[ii] = str[ii];
}
istr[ilen] = '\0';
if (*istr == '\0') { /* empty string input, not legal in FORTRAN */
ostr = (char*) emalloc(strlen("char(0)") + 1);
strcpy(ostr, "char(0)");
free(istr0);
return ostr;
}
ostr = cp = (char *) emalloc(12*ilen);
*ostr = '\0';
if (isprint((unsigned char)*istr)) { /* handle first character in input */
*cp++ = '\'';
switch (*istr) {
case '\'':
*cp++ = '\'';
*cp++ = '\'';
break;
case '\\':
*cp++ = '\\';
*cp++ = '\\';
break;
default:
*cp++ = *istr;
break;
}
*cp = '\0';
was_print = 1;
} else {
sprintf(tstr, "char(%d)", (unsigned char)*istr);
strcat(cp, tstr);
cp += strlen(tstr);
was_print = 0;
}
istr++;
for(ii = 1; ii < ilen; ii++) { /* handle subsequent characters in input */
if (isprint((unsigned char)*istr)) {
if (! was_print) {
strcat(cp, "//'");
cp += 3;
}
switch (*istr) {
case '\'':
*cp++ = '\'';
*cp++ = '\'';
break;
case '\\':
*cp++ = '\\';
*cp++ = '\\';
break;
default:
*cp++ = *istr;
break;
}
*cp = '\0';
was_print = 1;
} else {
if (was_print) {
*cp++ = '\'';
*cp = '\0';
}
sprintf(tstr, "//char(%d)", (unsigned char)*istr);
strcat(cp, tstr);
cp += strlen(tstr);
was_print = 0;
}
istr++;
}
if (was_print)
*cp++ = '\'';
*cp = '\0';
free(istr0);
return ostr;
}
static void
cl_netcdf(void)
{
int stat = nc_close(ncid);
check_err(stat);
}
static void
cl_c(void)
{
cline(" stat = nc_close(ncid);");
cline(" check_err(stat,__LINE__,__FILE__);");
#ifndef vms
cline(" return 0;");
#else
cline(" return 1;");
#endif
cline("}");
}
/* Returns true if dimension used in at least one record variable,
otherwise false. This is an inefficient algorithm, but we don't call
it very often ... */
static int
used_in_rec_var(
int idim /* id of dimension */
) {
int ivar;
for (ivar = 0; ivar < nvars; ivar++) {
if (vars[ivar].ndims > 0 && vars[ivar].dims[0] == rec_dim) {
int jdim;
for (jdim = 0; jdim < vars[ivar].ndims; jdim++) {
if (vars[ivar].dims[jdim] == idim)
return 1;
}
}
}
return 0;
}
/* Return name for Fortran fill constant of specified type */
static const char *
f_fill_name(
nc_type type
)
{
switch(type) {
case NC_BYTE:
return "NF_FILL_BYTE";
case NC_CHAR:
return "NF_FILL_CHAR";
case NC_SHORT:
return "NF_FILL_SHORT";
case NC_INT:
return "NF_FILL_INT";
case NC_FLOAT:
return "NF_FILL_FLOAT";
case NC_DOUBLE:
return "NF_FILL_DOUBLE";
}
derror("f_fill_name: bad type code");
return 0;
}
/* Generate Fortran for cleaning up and closing file */
static void
cl_fortran(void)
{
int ivar;
int idim;
char stmnt[FORT_MAX_STMNT];
char s2[FORT_MAX_STMNT];
char*sp;
int have_rec_var = 0;
/* do we have any record variables? */
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
have_rec_var = 1;
break;
}
}
if (have_rec_var) {
fline(" ");
fline("* Write record variables");
sprintf(stmnt, "call writerecs(ncid,");
/* generate parameter list for subroutine to write record vars */
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
/* if a record variable, include id in parameter list */
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(s2, "%s_id,", v->lname);
strcat(stmnt, s2);
}
}
sp = strrchr(stmnt, ',');
if(sp != NULL) {
*sp = '\0';
}
strcat(stmnt, ")");
fline(stmnt);
}
fline(" ");
fline("iret = nf_close(ncid)");
fline("call check_err(iret)");
fline("end");
fline(" ");
if (have_rec_var) {
sprintf(stmnt, "subroutine writerecs(ncid,");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(s2, "%s_id,", v->lname);
strcat(stmnt, s2);
}
}
sp = strrchr(stmnt, ',');
if(sp != NULL) {
*sp = '\0';
}
strcat(stmnt, ")");
fline(stmnt);
fline(" ");
fline("* netCDF id");
fline("integer ncid");
fline("* variable ids");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(stmnt, "integer %s_id", v->lname);
fline(stmnt);
}
}
fline(" ");
fline("include 'netcdf.inc'");
/* create necessary declarations */
fline("* error status return");
fline("integer iret");
/* generate integer/parameter declarations for all dimensions
used in record variables, except record dimension. */
fline(" ");
fline("* netCDF dimension sizes for dimensions used with record
variables");
for (idim = 0; idim < ndims; idim++) {
/* if used in a record variable and not record dimension */
if (used_in_rec_var(idim) && dims[idim].size != NC_UNLIMITED) {
sprintf(stmnt, "integer %s_len", dims[idim].lname);
fline(stmnt);
sprintf(stmnt, "parameter (%s_len = %lu)",
dims[idim].lname, (unsigned long) dims[idim].size);
fline(stmnt);
}
}
fline(" ");
fline("* rank (number of dimensions) for each variable");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(stmnt, "integer %s_rank", v->lname);
fline(stmnt);
}
}
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(stmnt, "parameter (%s_rank = %d)", v->lname,
v->ndims);
fline(stmnt);
}
}
fline("* starts and counts for array sections of record variables");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
sprintf(stmnt,
"integer %s_start(%s_rank), %s_count(%s_rank)",
v->lname, v->lname, v->lname, v->lname);
fline(stmnt);
}
}
fline(" ");
fline("* data variables");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim) {
char *sp;
fline(" ");
sprintf(stmnt, "integer %s_nr", v->lname);
fline(stmnt);
if (v->nrecs > 0) {
sprintf(stmnt, "parameter (%s_nr = %lu)",
v->lname, (unsigned long) v->nrecs);
} else {
sprintf(stmnt, "parameter (%s_nr = 1)",
v->lname);
}
fline(stmnt);
if (v->type != NC_CHAR) {
sprintf(stmnt, "%s %s(", ncftype(v->type),
v->lname);
/* reverse dimensions for FORTRAN */
for (idim = v->ndims-1; idim >= 0; idim--) {
if(v->dims[idim] == rec_dim) {
sprintf(s2, "%s_nr, ", v->lname);
} else {
sprintf(s2, "%s_len, ",
dims[v->dims[idim]].lname);
}
strcat(stmnt, s2);
}
sp = strrchr(stmnt, ',');
if(sp != NULL) {
*sp = '\0';
}
strcat(stmnt, ")");
fline(stmnt);
}
}
}
fline(" ");
/* Emit DATA statements after declarations, because f2c on Linux can't
handle interspersing them */
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
if (v->ndims > 0 && v->dims[0] == rec_dim && v->type != NC_CHAR) {
if (v->has_data) {
fline(v->data_stmnt);
} else { /* generate data statement for FILL
record */
size_t rec_len = 1;
for (idim = 1; idim < v->ndims; idim++) {
rec_len *= dims[v->dims[idim]].size;
}
sprintf(stmnt,"data %s /%lu * %s/", v->lname,
(unsigned long) rec_len,
f_fill_name(v->type));
fline(stmnt);
}
}
}
fline(" ");
for (ivar = 0; ivar < nvars; ivar++) {
struct vars *v = &vars[ivar];
/* if a record variable, declare starts and counts */
if (v->ndims > 0 && v->dims[0] == rec_dim) {
if (!v->has_data)
continue;
sprintf(stmnt, "* store %s", v->name);
fline(stmnt);
for (idim = 0; idim < v->ndims; idim++) {
sprintf(stmnt, "%s_start(%d) = 1", v->lname, idim+1);
fline(stmnt);
}
for (idim = v->ndims-1; idim > 0; idim--) {
sprintf(stmnt, "%s_count(%d) = %s_len", v->lname,
v->ndims - idim, dims[v->dims[idim]].lname);
fline(stmnt);
}
sprintf(stmnt, "%s_count(%d) = %s_nr", v->lname,
v->ndims, v->lname);
fline(stmnt);
if (v->type != NC_CHAR) {
sprintf(stmnt,
"iret = nf_put_vara_%s(ncid, %s_id, %s_start,
%s_count, %s)",
nfftype(v->type), v->lname, v->lname, v->lname,
v->lname);
} else {
sprintf(stmnt,
"iret = nf_put_vara_%s(ncid, %s_id, %s_start,
%s_count, %s)",
nfftype(v->type), v->lname, v->lname, v->lname,
v->data_stmnt);
}
fline(stmnt);
fline("call check_err(iret)");
}
}
fline(" ");
fline("end");
fline(" ");
}
fline("subroutine check_err(iret)");
fline("integer iret");
fline("include 'netcdf.inc'");
fline("if (iret .ne. NF_NOERR) then");
fline("print *, nf_strerror(iret)");
fline("stop");
fline("endif");
fline("end");
}
/* invoke netcdf calls (or generate C or Fortran code) to create netcdf
* from in-memory structure. */
void
define_netcdf(
const char *netcdfname)
{
char *filename; /* output file name */
if (netcdf_name) { /* name given on command line */
filename = netcdf_name;
} else { /* construct name from CDL name */
filename = (char *) emalloc(strlen(netcdfname) + 5);
(void) strcpy(filename,netcdfname);
if (netcdf_flag == -1)
(void) strcat(filename,".cdf"); /* old, deprecated extension */
else
(void) strcat(filename,".nc"); /* new, favored extension */
}
if (netcdf_flag)
gen_netcdf(filename); /* create netcdf */
if (c_flag) /* create C code to create netcdf */
gen_c(filename);
if (fortran_flag) /* create Fortran code to create netcdf */
gen_fortran(filename);
free(filename);
}
void
close_netcdf(void)
{
if (netcdf_flag)
cl_netcdf(); /* close netcdf */
if (c_flag) /* create C code to close netcdf */
cl_c();
if (fortran_flag) /* create Fortran code to close netcdf */
cl_fortran();
}
void
check_err(int stat) {
if (stat != NC_NOERR) {
fprintf(stderr, "ncgen: %s\n", nc_strerror(stat));
derror_count++;
}
}
/*
* For logging error conditions.
*/
#ifndef NO_STDARG
void
derror(const char *fmt, ...)
#else
/*VARARGS1*/
void
derror(fmt, va_alist)
const char *fmt ; /* error-message printf-style format */
va_dcl /* variable number of error args, if any */
#endif /* !NO_STDARG */
{
va_list args ;
if (lineno == 1)
(void) fprintf(stderr,"%s: %s: ", progname, cdlname);
else
(void) fprintf(stderr,"%s: %s line %d: ", progname, cdlname, lineno);
#ifndef NO_STDARG
va_start(args ,fmt) ;
#else
va_start(args) ;
#endif /* !NO_STDARG */
(void) vfprintf(stderr,fmt,args) ;
va_end(args) ;
(void) fputc('\n',stderr) ;
(void) fflush(stderr); /* to ensure log files are current */
derror_count++;
}
void *
emalloc ( /* check return from malloc */
size_t size)
{
void *p;
p = (void *) malloc (size);
if (p == 0) {
derror ("out of memory\n");
exit(3);
}
return p;
}
void *
ecalloc ( /* check return from calloc */
size_t size)
{
void *p;
p = (void *) calloc (size, 1);
if (p == 0) {
derror ("out of memory\n");
exit(3);
}
return p;
}
void *
erealloc ( /* check return from realloc */
void *ptr,
size_t size) /* if 0, this is really a free */
{
void *p;
p = (void *) realloc (ptr, size);
if (p == 0 && size != 0) {
derror ("out of memory");
exit(3);
}
return p;
}
/*
* For generated Fortran, change 'e' to 'd' in exponent of double precision
* constants.
*/
void
expe2d(
char *cp) /* string containing double constant */
{
char *expchar = strrchr(cp,'e');
if (expchar) {
*expchar = 'd';
}
}
/* Returns non-zero if n is a power of 2, 0 otherwise */
static
int
pow2(
int n)
{
int m = n;
int p = 1;
while (m > 0) {
m /= 2;
p *= 2;
}
return p == 2*n;
}
/*
* Grow an integer array as necessary.
*
* Assumption: nar never incremented by more than 1 from last call.
*
* Makes sure an array is within a factor of 2 of the size needed.
*
* Make sure *arpp points to enough space to hold nar integers. If not big
* enough, malloc more space, copy over existing stuff, free old. When
* called for first time, *arpp assumed to be uninitialized.
*/
void
grow_iarray(
int nar, /* array must be at least this big */
int **arpp) /* address of start of int array */
{
if (nar == 0) {
*arpp = (int *) emalloc(1 * sizeof(int));
return;
}
if (! pow2(nar)) /* return unless nar is a power of two */
return;
*arpp = (int *) erealloc(*arpp, 2 * nar * sizeof(int));
}
/*
* Grow an array of variables as necessary.
*
* Assumption: nar never incremented by more than 1 from last call.
*
* Makes sure array is within a factor of 2 of the size needed.
*
* Make sure *arpp points to enough space to hold nar variables. If not big
* enough, malloc more space, copy over existing stuff, free old. When
* called for first time, *arpp assumed to be uninitialized.
*/
void
grow_varray(
int nar, /* array must be at least this big */
struct vars **arpp) /* address of start of var array */
{
if (nar == 0) {
*arpp = (struct vars *) emalloc(1 * sizeof(struct vars));
return;
}
if (! pow2(nar)) /* return unless nar is a power of two */
return;
*arpp = (struct vars *) erealloc(*arpp, 2 * nar * sizeof(struct vars));
}
/*
* Grow an array of dimensions as necessary.
*
* Assumption: nar never incremented by more than 1 from last call.
*
* Makes sure array is within a factor of 2 of the size needed.
*
* Make sure *arpp points to enough space to hold nar dimensions. If not big
* enough, malloc more space, copy over existing stuff, free old. When
* called for first time, *arpp assumed to be uninitialized.
*/
void
grow_darray(
int nar, /* array must be at least this big */
struct dims **arpp) /* address of start of var array */
{
if (nar == 0) {
*arpp = (struct dims *) emalloc(1 * sizeof(struct dims));
return;
}
if (! pow2(nar)) /* return unless nar is a power of two */
return;
*arpp = (struct dims *) erealloc(*arpp, 2 * nar * sizeof(struct dims));
}
/*
* Grow an array of attributes as necessary.
*
* Assumption: nar never incremented by more than 1 from last call.
*
* Makes sure array is within a factor of 2 of the size needed.
*
* Make sure *arpp points to enough space to hold nar attributes. If not big
* enough, malloc more space, copy over existing stuff, free old. When
* called for first time, *arpp assumed to be uninitialized.
*/
void
grow_aarray(
int nar, /* array must be at least this big */
struct atts **arpp) /* address of start of var array */
{
if (nar == 0) {
*arpp = (struct atts *) emalloc(1 * sizeof(struct atts));
return;
}
if (! pow2(nar)) /* return unless nar is a power of two */
return;
*arpp = (struct atts *) erealloc(*arpp, 2 * nar * sizeof(struct atts));
}
#ifndef HAVE_STRLCAT
/* $OpenBSD: strlcat.c,v 1.12 2005/03/30 20:13:52 otto Exp $ */
/*
* Copyright (c) 1998 Todd C. Miller <address@hidden>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
/*
* Appends src to string dst of size siz (unlike strncat, siz is the
* full size of dst, not space left). At most siz-1 characters
* will be copied. Always NUL terminates (unless siz <= strlen(dst)).
* Returns strlen(src) + MIN(siz, strlen(initial dst)).
* If retval >= siz, truncation occurred.
*/
size_t
strlcat(char *dst, const char *src, size_t siz)
{
char *d = dst;
const char *s = src;
size_t n = siz;
size_t dlen;
/* Find the end of dst and adjust bytes left but don't go past end */
while (n-- != 0 && *d != '\0')
d++;
dlen = d - dst;
n = siz - dlen;
if (n == 0)
return(dlen + strlen(s));
while (*s != '\0') {
if (n != 1) {
*d++ = *s;
n--;
}
s++;
}
*d = '\0';
return(dlen + (s - src)); /* count does not include NUL */
}
#endif /* ! HAVE_STRLCAT */
/*
* Replace special chars in name so it can be used in C and Fortran
* variable names without causing syntax errors. Here we just replace
* each "-" in a name with "_MINUS_", each "." with "_PERIOD_", etc.
* For bytes with high bit set, from UTF-8 encoding of Unicode, just
* replace with "_xHH", where each H is the appropriate hex digit. If
* a name begins with a number N, such as "4LFTX", replace with
* "DIGIT_N_", such as "DIGIT_4_LFTX".
*
* Returned name is malloc'ed, so caller is responsible for freeing it.
*/
extern char*
decodify (
const char *name)
{
int count; /* number chars in newname */
char *newname;
const char *cp;
char *sp;
static int init = 0;
static char* repls[256]; /* replacement string for each char */
static int lens[256]; /* lengths of replacement strings */
static struct {
char c;
char *s;
} ctable[] = {
{' ', "_SPACE_"},
{'!', "_EXCLAMATION_"},
{'"', "_QUOTATION_"},
{'#', "_HASH_"},
{'$', "_DOLLAR_"},
{'%', "_PERCENT_"},
{'&', "_AMPERSAND_"},
{'\'', "_APOSTROPHE_"},
{'(', "_LEFTPAREN_"},
{')', "_RIGHTPAREN_"},
{'*', "_ASTERISK_"},
{'+', "_PLUS_"},
{',', "_COMMA_"},
{'-', "_MINUS_"},
{'.', "_PERIOD_"},
{':', "_COLON_"},
{';', "_SEMICOLON_"},
{'<', "_LESSTHAN_"},
{'=', "_EQUALS_"},
{'>', "_GREATERTHAN_"},
{'?', "_QUESTION_"},
{'@', "_ATSIGN_"},
{'[', "_LEFTBRACKET_"},
{'\\', "_BACKSLASH_"},
{']', "_RIGHTBRACKET_"},
{'^', "_CIRCUMFLEX_"},
{'`', "_BACKQUOTE_"},
{'{', "_LEFTCURLY_"},
{'|', "_VERTICALBAR_"},
{'}', "_RIGHTCURLY_"},
{'~', "_TILDE_"},
{'/', "_SLASH_"} /* should not occur in names */
/* {'_', "_UNDERSCORE_"} */
};
static int idtlen;
static int hexlen;
int nctable = (sizeof(ctable))/(sizeof(ctable[0]));
int newlen;
idtlen = strlen("DIGIT_n_"); /* initial digit template */
hexlen = strlen("_XHH"); /* template for hex of non-ASCII bytes */
if(init == 0) {
int i;
char *rp;
for(i = 0; i < 128; i++) {
rp = emalloc(2);
rp[0] = i;
rp[1] = '\0';
repls[i] = rp;
}
for(i=0; i < nctable; i++) {
size_t j = ctable[i].c;
free(repls[j]);
repls[j] = ctable[i].s;
}
for(i = 128; i < 256; i++) {
rp = emalloc(hexlen);
snprintf(rp, hexlen, "_X%2.2X", i);
rp[hexlen - 1] = '\0';
repls[i] = rp;
}
for(i = 0; i < 256; i++) {
lens[i] = strlen(repls[i]);
}
init = 1; /* only do this initialization once */
}
count = 0;
cp = name;
while(*cp != '\0') { /* get number of extra bytes for newname */
size_t j;
if(*cp < 0) { /* handle signed or unsigned chars */
j = *cp + 256;
} else {
j = *cp;
}
count += lens[j] - 1;
cp++;
}
cp = name;
if('0' <= *cp && *cp <= '9') { /* names that begin with a digit */
count += idtlen - 1;
}
newlen = strlen(name) + count + 1; /* bytes left to be filled */
newname = (char *) emalloc(newlen);
sp = newname;
if('0' <= *cp && *cp <= '9') { /* handle initial digit, if any */
snprintf(sp, newlen, "DIGIT_%c_", *cp);
sp += idtlen;
newlen -= idtlen;
cp++;
}
*sp = '\0';
while(*cp != '\0') { /* copy name to newname, replacing special chars */
size_t j, len;
/* cp is current position in name, sp is current position in newname */
if(*cp < 0) { /* j is table index for character *cp */
j = *cp + 256;
} else {
j = *cp;
}
len = strlcat(sp, repls[j], newlen);
assert(len < newlen);
sp += lens[j];
newlen -= lens[j];
cp++;
}
return newname;
}
/*
* Replace escaped chars in CDL representation of name such as
* 'abc\:def\ gh\\i' with unescaped version, such as 'abc:def gh\i'.
*/
void
deescapify (char *name)
{
const char *cp = name;
char *sp;
size_t len = strlen(name);
char *newname;
if(strchr(name, '\\') == NULL)
return;
newname = (char *) emalloc(len + 1);
cp = name;
sp = newname;
while(*cp != '\0') { /* delete '\' chars, except change '\\' to '\' */
switch (*cp) {
case '\\':
if(*(cp+1) == '\\') {
*sp++ = '\\';
cp++;
}
break;
default:
*sp++ = *cp;
break;
}
cp++;
}
*sp = '\0';
/* assert(strlen(newname) <= strlen(name)); */
strncpy(name, newname, len);
free(newname);
return;
}
