mathog at seqaxp.bio.caltech.edu wrote:
>> Any progress on a version of SRS 5.x that is ANSI C conformant (and so can
> be reasonably expected to work on nonUnix systems)?
>
the next version 5.0.6 will be - again - ANSI C conformant. We expect it
to be ready
about the end of July. The problem at the moment is the C function
library that
implements the Icarus commands. All these functions lack proper
prototypes but
they are automatically generated in the new version.
> When I tried to build SRS 5.0.5 with DEC C (on OpenVMS, but DEC C on DU
> would have flagged the same problems), enforcing the ANSI C standard, I ran
> into a bunch of type conflicts, missing prototypes, and so forth. I worked
> on cleaning it up for a while, but in the end, gave up when I encountered
> sections of code that are apparently only meaningful to C++ compilers. Ie,
> the
>> #ifndef _SDLoOBJ
> typedef struct SDLoOBJ *dummytointroducestructtagSDLoOBJ;
> #endif
These 'dummies' were also in SRS4 which compiled on openVMS. the problem
is that
a prototype with an unknown tag, eg, SDLoOBJ like
struct SDLoOBJ *myfunct (void);
works, however
void SDLoOBJ myfunct (struct SDLoOBJ*);
gives me warnings about local type definitions, the dummy makes sure
that the tag SDLoOBJ is known before the prototype.
...any suggestions to improve that?
regards
thure
