Bug#832468: bgpdump: FTBFS on s390x, ppc64, sparc64

2016-07-28 Thread Christoph Biedl
tags 832468 help
thanks

Christoph Biedl wrote...

> The bgpdump fails to build on big-endian archs where 
> 
> sizeof(time_t) == sizeof(long) == sizeof(void *) == 8

There's a fix in my queue but I want to test it before upload. If
anybody can grant me access to a box with any of the above
architectures, please drop me a line.

Else I'll have no choice but to drop s390x from the list of
architectures this package is built for.

Christoph


signature.asc
Description: Digital signature


Bug#832468: bgpdump: FTBFS on s390x, ppc64, sparc64

2016-07-25 Thread Christoph Biedl
Package: bgpdump
Version: 1.4.99.15+hg127-2
Severity: important

The bgpdump fails to build on big-endian archs where 

sizeof(time_t) == sizeof(long) == sizeof(void *) == 8

Reason is the loader for 32 bit timestamps in bgpdump_lib.c:144 that
stuffs the data into the first bytes (but is endianness aware). For LE
this works as long as the remaining bytes are zero. For BE this
creates insanely huge time values where "It may return NULL when the
year does not fit into an integer" (gmtime manpage) applies. Things go
downhill from there.

Christoph


signature.asc
Description: Digital signature