[Xapian-discuss] xapian-1.0.12 compiled error with debug mode

Olly Betts olly at survex.com
Thu May 14 03:16:18 BST 2009


On Wed, May 13, 2009 at 01:12:13PM +0800, ?????? wrote:
> Hello, for some reason, my xapian often have follow errors, I know this means
> xapian run out of memory...
> 
> --->>
> terminate called after throwing an instance of 'std::bad_alloc'
>   what():  St9bad_alloc
> Abort trap: 6 (core dumped)
> <<---

The most useful way to find out what's happening is probably just to
poke the core dump with gdb ("gdb core" and "bt" to given a backtrace
should do it.

That may just tell you it was in terminate() called by
the runtime code though.  You might get a more useful backtrace by
running it under gdb: put "gdb --args" in front of the command to run
your program, then "run" and "bt" to produce a backtrace.

> So, I try to re-compile the xapian with debug mode (add --enable-log=yes
> --enable-assertions=yes to configure options), but got the folowing error
> messages:

Oops, the first is already fixed on trunk, but we failed to backport the
fix.  Thanks for letting us know.  It will be fixed for 1.0.13 like so:

http://xapian.org/C?12707?branches/1.0/xapian-core/api/omdatabase.cc

> backends/flint/flint_table.cc: In constructor `FlintTable::FlintTable(std::string,
> bool, int, bool)':
> backends/flint/flint_table.cc:1566: error: `tablename_' undeclared (first use this
> function)

This one was just badly backported (there's a new parameter here on
trunk) - fix is:

http://xapian.org/C?12708?branches/1.0/xapian-core/backends/flint/flint_table.cc

Cheers,
    Olly



More information about the Xapian-discuss mailing list