p.s. nevertheless, some temporary data storage could be introduced for
large tables. nothing formatted, nothing spectacular, nothing "final"
just a plain dumb temporary storage in case of large table dumps. will
do something :)
kr
On Thu, Aug 18, 2011 at 10:09 AM, Miroslav Stampar
wrote:
> hi.
hi.
there was an important fix (r4354) related to this same problem.
there was really a problem going on where connection dropping was
causing loss of all data.
the problem was reintroduced when I've done some (single and multi
threading generic) refactoring a month or two ago. my fault. sorry.
