On 7/9/07, Francesco Romani <[EMAIL PROTECTED]> wrote:
> ,---- > | ../pvm3/pvm_parser.c: In function 'f_skip_space': > | ../pvm3/pvm_parser.c:48: warning: implicit declaration of function 'cf_skip_frontwhite' > | ../pvm3/pvm_parser.c:48: warning: assignment makes pointer from integer without a cast > | ../pvm3/pvm_parser.c: In function 'f_pvm_parser': > | ../pvm3/pvm_parser.c:57: error: 'CF_ROOT_TYPE' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:57: error: (Each undeclared identifier is reported only once > | ../pvm3/pvm_parser.c:57: error: for each function it appears in.) > | ../pvm3/pvm_parser.c:57: error: 'p_root' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:58: error: 'CF_SECTION_TYPE' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:58: error: 'p_section' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:59: error: 'CF_SUBSECTION_TYPE' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:59: error: 'p_subsection' undeclared (first use in this function) > | ../pvm3/pvm_parser.c:60: error: 'CF_KEYVALUE_TYPE' undeclared (first use in this function) > `----
[...]
I'll take a look ASAP. Until this get fixed, I recommend to disable pvm module in 1.1.0 builds, since it can be already broken.
After a first glance, looks like the main problem in here is to rewrite the pvm.cfg parser. Not that hard, but somewhat boring. Maybe our brand new cfg parser needs some extensions too to handle subsections in configuration files (not yet checked) or maybe is just simpler to slightly change the pvm.cfg syntax (this is nasty but will also reflect the fact that pvm3 support is expected to NOT work flawlessly on 1.1.x) I'll address this task once new decoder code landed on CVS HEAD, and we're pretty near to this event, except for obvious debug/testing/docs I'm just missing some probing details. Bests, -- Francesco Romani // Ikitt
