I don't quite understand the docs, but

送交者: ak70 于 2010-09-09, 05:25:58:

回答: also, if i use v[] directly, no error happens 由 qtl 于 2010-09-09, 05:18:02:

n
INTEGER. Specifies the number of elements in vector x.
x
REAL for sasum
DOUBLE PRECISION for dasum
COMPLEX for scasum
DOUBLE COMPLEX for dzasum
Array, DIMENSION at least (1 + (n-1)*abs(incx)).
incx
INTEGER. Specifies the increment for indexing vector x.

why do you use &n,&inc where, according to the doc, they should be integers? Is this some special calling convention?

所有跟贴:

yes, to call fortran with c/c++ - qtl (45 bytes) 2010-09-09, 05:30:27 (484703)
- then maybe you should try &v instead of v. (无内容) - ak70 (0 bytes) 2010-09-09, 05:36:21 (484704)
  - v is the initial address already for the vector (无内容) - qtl (0 bytes) 2010-09-09, 05:44:33 (484705)
    - yes, but maybe you should try to use the pointer to pointers - ak70 (82 bytes) 2010-09-09, 05:49:23 (484706)
      - your this suggestion is wrong anyway - qtl (40 bytes) 2010-09-09, 05:51:39 (484707)
        
        if use &v, then **v will be passed - qtl (25 bytes) 2010-09-09, 06:01:24 (484710)
        
        what was the problem? (无内容) - ak70 (0 bytes) 2010-09-09, 05:59:29 (484709)

加跟贴