Re: Lua 5.2.0 compile error

[Geoff Down <geoffdown@...>]

On Sun, Nov 2, 2014, at 03:28 PM, Roberto Ierusalimschy wrote:
> >  Yes, that's one way for the end-user to fix it - but they may not have
> >  the knowledge even to know where to begin looking for the problem.
> > If it gets fixed in liblua, then everything that uses liblua
> > automatically starts working again as soon as the new version propagates
> > through the dependent software.
> 
> At the time of the original message, I replied this:
> 
>   This seems weird. Both fseeko and ftello use the type 'off_t' in their
>   prototypes. If the compiler did not complain when reading the header
>   files, either their prototypes are missing from 'stdio.h' (where these
>   functions should be declared) or their prototypes are not using this
>   type. Both cases seem to break POSIX compliance.

Actually in stdio.h:
#if !defined(_ANSI_SOURCE) && !defined(__STRICT_ANSI__)
typedef __darwin_off_t          fpos_t;
#else
typedef __int64_t               fpos_t;
#endif
...
int      fseeko(FILE *, fpos_t, int);
fpos_t   ftello(FILE *);

So there is a native __darwin_off_t (which sys/types.h aliases to
off_t).


> 
> I still stand for what I said. You should complain with them, not with
> us. If it gets fixed there, then every software that asssumes correct
> headers will benefit.
> 
> -- Roberto
> 
 You are the fourth person to say to me 'it's someone else's fault, get
 them to fix it'. Nothing personal, but I'm frankly sick of it.
If it's within your power to fix it, why stand on ceremony? You already
make allowances for different compile environments, this is just one
more.
I have the source files for 'bash' here: 'configure' checks whether
off_t is a valid type and if not equates it to 'long'.
I don't even know who this 'them' is. If it's Apple, they aren't going
to give a rat's - unsupported OS. So there is no mechanism to get the
updated stdio.h onto people's computers.
 People generally just want their car to work, they don't want to have
 to find the geologist that surveyed the wrong oil deposits that made
 the wrong sort of petrol that clogged the particular carburetor
 that....etc.
GD

-- 
http://www.fastmail.fm - IMAP accessible web-mail