Thanks for this, it not only fixes a memory leak but also a regression.
with the previous version of the fix after the processing of the first
tombstone nstombstone_vals was set an duse in deciding what to index :-)
On 08/28/2015 02:19 AM, Noriko Hosoi wrote:
Description: Commit ca4d38dff3bd820af2bf60c9bcc82fd64aac2556 fixing
ticket #48252 introduced memory leaks.
. nstombstone_vals needs to be allocated just once and freed at
. when skipping to index an entry for being a tombstone, the entry
to be freed.
389-devel mailing list