Re: [platform-testers] inetutils-2.5.90 released [alpha]

[Bruno Haible]

Hi Collin,

> process.c:224:50: warning: argument to 'sizeof' in 'strncmp' call is the same 
> expression as the first source; did you mean to provide an explicit length? 
> [-Wsizeof-pointer-memaccess]
>   224 |       if (!strncmp (UT_USER (uptr), name, sizeof (UT_USER (uptr))))
>       |                                                  ^
> 
> And again:
> 
> syslogd.c:1648:43: warning: argument to 'sizeof' in 'strncpy' call is the 
> same expression as the source; did you mean to provide an explicit length? 
> [-Wsizeof-pointer-memaccess]
>  1648 |       strncpy (line, utp->ut_line, sizeof (utp->ut_line));
>       |                                           ^
> syslogd.c:1666:30: warning: argument to 'sizeof' in 'strncmp' call is the 
> same expression as the second source; did you mean to provide an explicit 
> length? [-Wsizeof-pointer-memaccess]
>  1666 |                       sizeof (UT_USER (utp))))
>       |                              ^
> 
> I've been meaning to fix these after noticing them recently, but I'm not
> very familiar with readutmp. Perhaps this code was for ancient versions
> of the module?

Definitely, yes. The gnulib/NEWS entries:

2023-08-08  readutmp        The result element type of the function read_utmp,
                            STRUCT_UTMP, is no longer the same as the result
                            value type of the function getutxent, struct utmpx.

2023-08-03  readutmp        Some STRUCT_UTMP members can be char *,
2023-08-01                  rather than fixed-length char arrays.
                            On some platforms, the timestamp is ut_ts of type
                            struct timespec, not ut_tv of type struct timeval,
                            and ut_session is pid_t not long.
                            read_utmp's 2nd arg is now idx_t * not size_t *.
                            Link additionally with $(READUTMP_LIB).

Basically, before these changes, using readutmp implied platform-dependent
code and #ifs. Now, the API is the same on all platforms, no #ifs needed any
more. But it requires some code changes (mostly simplifications) on the
inetutils side.

Bruno