mirror of https://github.com/BOINC/boinc.git
3983504893
fix was incomplete. In order for it to work as intended, I would have also had to modify SCHED_RESULT_ITEM::parse to copy the additional needed fields into the in-memory structure. But this is fragile. The next time some additional fields are added to the result table, they would have to be incorporated here as well, to ensure that SET followed by UPDATE is the identity operation. So I did a more graceful and robust fix. Simply set result.id=0 for those results that have already been received or which for other reasons should not be modified in the database, then skip these when updating. svn path=/trunk/boinc/; revision=4763 |
||
---|---|---|
.. | ||
.cvsignore | ||
Makefile.am | ||
boinc_db.C | ||
boinc_db.h | ||
constraints.sql | ||
db_base.C | ||
db_base.h | ||
init_db | ||
mysql.h | ||
schema.sql |