Martin Michlmayr <[EMAIL PROTECTED]> writes:
> Package: remctl
> Version: 2.1-1
> Severity: serious
> This package no longer builds in unstable. Below is a patch that
> addresses the error but it's probably not optimal since it introduces
> some warnings.
>> tests/runtests.c: In function 'test_summarize':
>> tests/runtests.c:472: error: assignment of read-only member '__in'
> --- tests/runtests.c~ 2006-09-07 12:27:10.223337750 +0200
> +++ tests/runtests.c 2006-09-07 12:30:20.123205750 +0200
> @@ -469,7 +469,7 @@
> printf(" (exit status %d)", status);
> else if (status < 0)
> printf(" (killed by signal %d%s)", -status,
> - WCOREDUMP(ts->status) ? ", core dumped" : "");
> + WCOREDUMP((struct testset *)ts->status) ? ", core dumped" :
> "");
> putchar('\n');
> }
That is a bizarre error message and an even more bizarre patch to fix it.
The WCOREDUMP macro should take an integer exit status and return true or
false based on whether the process exited with a core dump. Passing a
pointer into it is definitely wrong; casting it to a pointer to some
struct that WCOREDUMP has never heard about is even weirder. Something
very, very strange is going on here.
That looks suspiciously like some sort of bug in the libc headers.
However, that's almost never the actual problem, even when it looks like
it, so I'll take a closer look.
Thank you for the report!
--
Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/>
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
