lp:~jameinel/runsnakerun/memory_dump_integration
- Get this branch:
- bzr branch lp:~jameinel/runsnakerun/memory_dump_integration
Branch merges
Branch information
Recent revisions
- 53. By John A Meinel
-
/call counts now use 1.0, as it has the most meaning.
Also, start calling 'remove_expensive_ references' , since it seems to
give cleaner output. - 51. By John A Meinel
-
Include values since it helps a bit.
Also, recursive is now overloaded as the number of references.
(Calls is the number of referrers.)
Also, we now compute *something* for child_cummulative_time( ), which
means that we actually get results in the squaremap. - 50. By John A Meinel
-
Some updates now that we have a way to 'compute_
total_size' .
Note, however, that this is rather expensive. As in a minute or so to
update from just thousands of references.
The problem is probably the number of entries that end up in the global
refcycle by virtue of some sort of module access. - 49. By John A Meinel
-
Initial attempt at wrapping the memory dump output into the graph that runsnakerun wants.
Currently the breakdown is done on every object....
It would probably be better to aggregate things by type. - 48. By John A Meinel
-
Change the 'already_done' dict into an 'already_done' set.
Should save some memory when doing large recursions.
Branch metadata
- Branch format:
- Branch format 6
- Repository format:
- Bazaar RepositoryFormatKnitPack6 (bzr 1.9)