Repository navigation
_tracemalloc: remove pointer_t type #84789
Description
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee = None
closed_at = <Date 2020-05-14.01:12:44.737>
created_at = <Date 2020-05-12.22:47:15.350>
labels = ['library', '3.9']
title = '_tracemalloc: remove pointer_t type'
updated_at = <Date 2020-05-14.01:12:44.735>
user = 'https://github.com/vstinner'bugs.python.org fields:
activity = <Date 2020-05-14.01:12:44.735>
actor = 'vstinner'
assignee = 'none'
closed = True
closed_date = <Date 2020-05-14.01:12:44.737>
closer = 'vstinner'
components = ['Library (Lib)']
creation = <Date 2020-05-12.22:47:15.350>
creator = 'vstinner'
dependencies = []
files = []
hgrepos = []
issue_num = 40609
keywords = ['patch']
message_count = 7.0
messages = ['368747', '368750', '368753', '368754', '368755', '368757', '368812']
nosy_count = 1.0
nosy_names = ['vstinner']
pr_nums = ['20059', '20060', '20062', '20064', '20065']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue40609'
versions = ['Python 3.9']
Activity
vstinner commentedon May 12, 2020
In bpo-26588, I modified the _tracemalloc module to support tracing memory allocations in multiple "domains" for numpy. I modified hashtable.c to support keys larger than void*, and I modified _tracemalloc.c to use a new pointer_t structure which is made of (domain: unsigned int, ptr: void*).
These changes made hashtable.c way more complicated than what it should be. I would like to reuse hashtable.c in more places. I am working on a change to remove the pointer_t type from _tracemalloc.c, and instead use a hash table for traces of domains other than the default domain (0). The new hash table maps a domain to a hash table: traces of a domain.
tracemalloc_traces remains the default hash table for traces of the default domain. Since it's the most common case, it doesn't go through the new hash table.
vstinner commentedon May 12, 2020
New changeset 9e2ca17 by Victor Stinner in branch 'master':
bpo-40609: Rewrite how _tracemalloc handles domains (GH-20059)
9e2ca17
vstinner commentedon May 13, 2020
New changeset f9b3b58 by Victor Stinner in branch 'master':
bpo-40609: Remove _Py_hashtable_t.key_size (GH-20060)
f9b3b58
vstinner commentedon May 13, 2020
New changeset 2d0a3d6 by Victor Stinner in branch 'master':
bpo-40609: Add destroy functions to _Py_hashtable (GH-20062)
2d0a3d6
vstinner commentedon May 13, 2020
New changeset d95bd42 by Victor Stinner in branch 'master':
bpo-40609: _tracemalloc allocates traces (GH-20064)
d95bd42
vstinner commentedon May 13, 2020
New changeset 5b0a303 by Victor Stinner in branch 'master':
bpo-40609: _Py_hashtable_t values become void* (GH-20065)
5b0a303
vstinner commentedon May 14, 2020
I modified _Py_hashtable_t to remove key_size and data_size: keys are now always "void *" and values are now always "void *".
See also bpo-40602: Move Modules/hashtable.h to Include/internal/pycore_hashtable.h.