archive-index uses too much memory
Bug #804144 reported by
Jjed
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| archive-index |
Confirmed
|
Low
|
Unassigned | ||
Bug Description
archive-index currently uses as much RAM as the size of an archive's uncompressed Contents file. It should store Contents paths in a tree data structure to decrease data redundancy.
This is also a roadblock for performance improvements decreasing the amount of regex searching archive-index performs.
| Changed in archive-index: | |
| importance: | High → Low |
| assignee: | Jacob Johan Edwards (j-johan-edwards) → nobody |
Revision history for this message
| Jjed (jjed) wrote | #1 |
To post a comment you must log in.
Unfortunately the memory overhead for Python objects is so high that a tree structure consumes even more memory than the current implementation, even using `___slots__`. Perhaps a C extension would be desirable.