compiling libdap-3.6.2 on Solaris 10 x86

Dale Castle dale at vims.edu
Wed Feb 7 04:44:32 PST 2007


James,

 	Thanks for your continued help with this. I get a warning early on 
in the make, but then GetOpt.cc won't compile.

cc: Warning: illegal option -features=tmplrefstatic

  CC -DHAVE_CONFIG_H -I. -I. -I. -I. -I./gl -I/usr/local/include 
-I/usr/include/libxml2 -I./GNU -xO3 -mt -fsimple=1 -ftrap=%none -nofstore 
-xbuiltin=%all -xlibmil -xlibmopt -xtarget=opteron -xregs=no%frameptr 
-features=tmplrefstatic -c GNU/GetOpt.cc -KPIC -DPIC -o .libs/GetOpt.o

"GNU/GetOpt.cc", line 70: Error: Too many arguments in call to "alloca()".
1 Error(s) detected.

This is near line 70:

   int nonopts_size
     = (last_nonopt - first_nonopt) * sizeof (char *);
   char **temp = (char **) alloca (nonopts_size);

That's about where the make dies:

make: Fatal error: Command failed for target `GetOpt.lo'
Current working directory /export/home/dale/libdap/libdap-3.6.2
make: Fatal error: Command failed for target `all-recursive'
Current working directory /export/home/dale/libdap/libdap-3.6.2
make: Fatal error: Command failed for target `all'

On Mon, 5 Feb 2007, James Gallagher wrote:

>
> On Feb 2, 2007, at 3:08 PM, Dale Castle wrote:
>
>> James,
>> 
>>    Thank you for the response. I tried that flag you suggested and it did 
>> not recognize that when I ran make.
>
> Do you mean the code didn't compile or the option was not recognized by the 
> compiler? I'm asking because the compiler error message from your original 
> post suggests using that option... Here's the part:
>>>> "Operators.h", line 80: Error: Reference to static dods_max(int,
>>>> int) not allowed in template static USCmp<unsigned char, short>::eq
>>>> (unsigned char, short), try using -features=tmplrefstatic.
>
>                                       ^^^^^^^^^^^^^^^^^^^^^^^
>
> Or maybe looking at the other values -features can take to see if there's 
> another option for templates. If not, try removing the word 'static' since 
> it's there to limit scope but I think the code will work just fine either 
> way.
>> 
>
> James
>
> --
> James Gallagher                jgallagher at opendap.org
> OPeNDAP, Inc                   406.723.8663
>
>



More information about the Opendap-tech mailing list