Glenn ha escrito:
JOAQUIN LOPEZ MU?Z wrote:
----- Mensaje original -----
... I guess there is a bug in boost/multi_index/sequenced_index.hpp. ...
This is indeed a bug, and a gross one I must embarrassingly admit :) Also, the proposed patch is the way to go. Thank you very much for spotting this, I'll be committing the fixes next Monday (during weekends I'm CVS-disabled.) FYI, the same problem also shows at other places like random-access indices and some other ops like merge, yet the reference docs use the correct expression &x==this everywhere, it's so disturbing, I wonder how this could go unnoticed for so long.
Well, ask yourself why your package's regression tests don't cover these cases -- and make sure they do, for all of whatever code changes you're about to make. This is critical for my own project; I was just about to commit to using multi_index, but now I'm not so sure of its reliability.
Well, if you think I can't meet the reliability standards that you mandate of the 3rd party SW you use, then by all means don't use the lib --luckily for you, this is *not* critical for your project since you haven't commited to using Boost.MultiIndex yet. Let me also state, however, that I find your reaction to this (fixed as of three days later) bug report quite bewildering; what kind of bug-free software are you using?
I take it this would mean that, at a minimum, I would be unable to use the 1.32.0 release already installed with the OS, and I would need to pull directly from CVS after your changes go in, at least until 1.34.0 shows up.
That's correct if you use or plan to use sequenced indices splice or merge. Otherwise, this particular problem does not affect you.
That makes me wonder as well: is there some way to post a warning on the Boost site that this package is broken in the current releases?
You'd have to convince one of the Boost developers with write access to the CVS that this issue is important enough that she enter the warning for you.
Glenn
Joaquín M López Muñoz Telefónica, Investigación y Desarrollo