Bug 1455346 - Update python-idna to version 2.5
Summary: Update python-idna to version 2.5
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: RDO
Classification: Community
Component: distribution
Version: trunk
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: Milestone2
: trunk
Assignee: Alan Pevec
QA Contact: Shai Revivo
URL:
Whiteboard:
Depends On:
Blocks: RDO-PIKE
TreeView+ depends on / blocked
 
Reported: 2017-05-24 21:03 UTC by Alan Pevec
Modified: 2017-06-02 13:20 UTC (History)
8 users (show)

Fixed In Version: python-idna-2.5-1.el7
Doc Type: Rebase: Bug Fixes and Enhancements
Doc Text:
Clone Of: 1455078
Environment:
Last Closed: 2017-06-02 13:20:42 UTC


Attachments (Terms of Use)

Description Alan Pevec 2017-05-24 21:03:59 UTC
+++ This bug was initially created as a clone of Bug #1455078 +++

Description of problem:

Currently python-idna-2.0 is shipped in RHEL 7.3. This version consumes around 25MB of memory only to load:

>>> import resource
>>> resource.getrusage(resource.RUSAGE_SELF).ru_maxrss
5052
>>> import idna
>>> resource.getrusage(resource.RUSAGE_SELF).ru_maxrss
31344
>>> 

Some memory optimizations have been applied upstream from release 2.0 until latest one, 2.5 leading to a huge improvement:

>>> import resource
>>> resource.getrusage(resource.RUSAGE_SELF).ru_maxrss
5048
>>> import idna
>>> resource.getrusage(resource.RUSAGE_SELF).ru_maxrss
5888
>>> 

Reviewing the commits from v2.0..v2.5, there have been no changes in the public methods of the library, so it seems to be fully compatible with version 2.0, so i'd say it should be safe to update it.


OpenStack services using eventlet are using idna via import from python-dns, which is leading to increase memory usage.

Comment 1 Alan Pevec 2017-05-24 22:12:14 UTC
rebuilding latest 2.5 from Fedora with a small patch to disable py3:
https://github.com/rdo-common/python-idna/commit/183eb7316a0e0687c7224103d75d04f351072511

Comment 2 Alan Pevec 2017-05-24 22:27:34 UTC
Tagged for common-pending repo, please test if it is safe for common-testing.

Comment 5 Alan Pevec 2017-05-25 13:42:20 UTC
in common-testing

Comment 6 Haïkel Guémar 2017-06-02 13:20:42 UTC
Tagged in common-release since


Note You need to log in before you can comment on or make changes to this bug.