1) Make memory counters aggregated.
2) Add attachment pool and allocate appropriate resources out of this pool.
3) Always release attachments explicitly (via destructor).
4) Always delete user requests prior to attachment deletion.
5) Introduce memory usage counters per every monitoring object.
6) Misc refactoring.
Some pieces are still incomplete (although everything basically works), but I'd like to get feedback and testing sooner rather than later.
Sixth step: remove the places protected by the macro PC_ENGINE. More to come. (I guess all of you who have pending functional changes to post will love this commit. <g>)
2.- style
3.- Try to fix the mess caused by different data types used to represent bifields of SCL_ constants defined in scl.h,
first round. No circular dependencies found through MSVC. Still pending dyn* and grant.epp files.
PandoraBox* pbox = reinterpret_cast<PandoraBox*>(&can_of_worms);
pbox->open();
pbox->flush();
Nickolay may want to undo my ods.h changes if gcc insists
in its crusade against non-PODs and poodles.
const correctness
BOOLEAN -> bool
Vars in scope
Formatting
JRD_REQ -> jrd_req*
JRD_TRA -> jrd_tra*
JRD_REL -> jrd_rel*
JRD_FLD -> jrd_fld*
event -> event_t, EVENT_T dropped, EVENT should go in the future
Frank's corrections
alice's STR -> alice_str*
remote's STR -> rem_str*
etc.
It creates entities and backups and restores for me. :-)
1. Bring trace DSQL, DYN and BLR trace logging up-to-date
2. Remove redundant CSB pointer dereferences in BLR parser and request compiler
3. Convert RIGHT JOIN to LEFT JOIN early during BLR parsing
4. Check BLR syntax for unititalized contexts usage
5. Some type-safety and const-correctness fixes
in other requests. Make request access lists use view ID, not pointer
to view structure. This changes solve the crash reported by
Evgeny Kilin when altering complex recursive procedure->trigger chains.
