Build failures in unit test when firebird is enabled

[Michael Stahl]

On 16.02.2016 22:22, Ken Moffat wrote:
> This is on 5.1.0.3, but I saw similar on 5.0 - at that time, I
> needed the app and I hadn't built gdb.  Found a few references, and
> people building with firebird disabled, looks like the problem hits
> some people from time to time.  This is on an i7 haswell.

yeah firebird tests have never been all that reliable...

> Thread 1 (Thread 0x2b4bfee4c700 (LWP 28212)):
[...]
> #6  0x00002b4bf3e8ea0a in abort () from /lib/libc.so.6
> No symbol table info available.
> #7  0x00002b4bf3ed1970 in malloc_printerr () from /lib/libc.so.6
> No symbol table info available.
> #8  0x00002b4bfb1dd51f in (anonymous namespace)::threadStart (arg=0x2b4bfd1f95d0) at ../src/jrd/ThreadStart.cpp:138
>         mainThreadContext = {savedPool = 0x0}
>         localArgs = {routine = 0x2b4bfb1e2670 <(anonymous namespace)::shutdownThread(void*)>, arg = 0x0}
> #9  0x00002b4bf4409404 in start_thread () from /lib/libpthread.so.0
> No symbol table info available.

so some crash inside of Firebird... this is an experimental feature
anyway, so either you try to debug it or use --disable-firebird-sdbc