Re: (ITS#7238) Multiple txns in same thread break mdb

[hyc@symas.com]

h.b.furuseth@usit.uio.no wrote:
>   Related doc bug:
>
>   I expect it should be "only one thread at a time".  I doubt
>   it's about pinning some operation to a particular thread:-)

Wrong.

For write txns I suppose that could be relaxed. For readers, the reader slot 
is set when the txn is created, and that slot is owned by a particular thread.

>   However, maybe it should rather be something like "No other
>   transaction doing mdb_open() may exist at the same time, nor
>   a write transaction (since that also updates the databases),
>   nor may such a transaction exist during mdb_close()" ?
>
>   I note mdb_drop() also updates the databases and can
>   likely not coexist with the others, but that's covered by
>   mentioning write txns.
>
>   Regarding mdb_drop(), while I'm at it:
>   	@brief Delete a database and/or free all its pages.
>   	@param[in] del non-zero to delete the DB from
>   	the environment, otherwise just free its pages.
>   Free which pages?  In the DB, or some malloced ones?

This is single-level-store, there is no distinction. In the DB or in memory, 
all the same.

-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/